CN104281516A

CN104281516A - Methods and apparatus to characterize households with media meter data

Info

Publication number: CN104281516A
Application number: CN201410341892.1A
Authority: CN
Inventors: B·尚卡; M·波派; T·多尔森; D·J·库尔泽尼斯基; J·加西亚; L·赫穆拉; H·尤; P·C·多伊; C·布尔奎因; J·M·布尔
Original assignee: Nielsen Co US LLC
Current assignee: Nielsen Co US LLC
Priority date: 2013-07-09
Filing date: 2014-07-09
Publication date: 2015-01-14
Also published as: HK1206122A1; EP2824854A1; CA2856246A1

Abstract

Methods, apparatus, systems and articles of manufacture are disclosed to characterize households with media meter data. An example method includes identifying, with a processor, a power status and a first automatic gain control (AGC) value for an exposure minute from a panelist audience meter in a first household, the panelist audience meter comprising a power sensor, identifying a second AGC value and a daypart for a household tuning minute from a first media meter (MM) in the first household, the MM comprising microphones to collect audio data, and calculating model coefficients based on the exposure minute and the household tuning minute to be applied to data from a second MM in a second household, the model coefficients to facilitate a power status probability calculation in the second household devoid of the panelist audience meter having the power sensor.

Description

The method and apparatus of family is characterized by media gage data

Related application

This application claims the U.S.Provisional Serial 61/839 submitted on June 25th, 2013, 344 (attorney 20004/104617US01), the U.S.Provisional Serial 61/844 that on July 9th, 2013 submits to, 301 (attorney 20004/104617US02), the U.S.Provisional Serial 61/986 that on April 30th, 2014 submits to, 409 (attorney 20004/104617US03), the U.S.Provisional Serial 62 that on June 4th, 2014 submits to, 007, the rights and interests of 535 (attorney 20004/104617US04), the full content of these applications is incorporated herein by reference all hereby.

Technical field

The disclosure relates generally to market survey, more specifically, relates to the method and apparatus characterizing family (household) by media gage data (media meter data).

Background technology

In recent years, the research of group member is made great efforts to be included in and is met one or more and pay close attention in demographic qualified family and install and measure hardware.In some cases, measure hardware can by lead to from media presentation device (such as, televisor) measuring instrument rigid line connect determine media presentation device whether be energized and be tuned to certain station.In other cases, measure hardware and can determine which kinsfolk is exposed to the specific part of media by pressing one or more button on people meter (People Meter) near the kinsfolk of TV.

Summary of the invention

A method for computing medium installation's power source shape probability of state, described method comprises:

Purpose processor identification is for the power supply status of the exposure minute from the group member's audience measurement instrument in First Family and the first automatic growth control (AGC) value, and described group member's audience measurement instrument comprises power sensor;

Identify for family's the 2nd AGC value and the period of tuning minute from the first media measuring instrument (MM) in described First Family, MM comprises the microphone for collecting voice data;

Within tuning minute, calculate the model coefficient of the data be applied to from the 2nd MM in the second family with described family based on described exposure minute, described model coefficient contributes to lacking the power supply status probability calculation in described second family of described group member's audience measurement instrument with described power sensor.

An equipment for computing medium installation's power source shape probability of state, described equipment comprises:

Automatic growth control (AGC) watch-dog, it identifies power supply status for the exposure minute from the group member's audience measurement instrument in First Family and an AGC value, described group member's audience measurement instrument comprises power sensor, the identification of AGC watch-dog is for family's the 2nd AGC value and the period of tuning minute from the first media measuring instrument (MM) in described First Family, and MM comprises the microphone for collecting voice data;

Modeling engine, it calculates the model coefficient of the data be applied to from the 2nd MM in the second family with described family based on described exposure minute for tuning minute, and described model coefficient contributes to lacking the power supply status probability calculation in described second family of described group member's audience measurement instrument with described power sensor.

Comprise a tangible machine readable storage medium storing program for executing for instruction, described instruction causes machine at least upon being performed:

Identify the power supply status for the exposure minute from the group member's audience measurement instrument in First Family and the first automatic growth control (AGC) value, described group member's audience measurement instrument comprises power sensor;

Identify for family's the 2nd AGC value and the time of tuning minute from the first media measuring instrument (MM) in described First Family, MM comprises the microphone for collecting voice data;

Accompanying drawing explanation

Fig. 1 illustrates that useable medium gage data characterizes the example media distributional environment of family.

Fig. 2 is the schematic diagram collecting engine according to the example of instruction structure of the present disclosure.

Fig. 3 is the figure line of the example viewing index effect illustrated based on the age of the data collected.

Fig. 4 is for the exemplary weights allocation table to minute Applicative time weight collected.

Fig. 5 is the example metrics subset map that the independent distribution characterizing family's tolerance that family uses by media gage data is shown.

Fig. 6 to Fig. 9 represents the process flow diagram that the example that can be performed to realize Fig. 1 and Fig. 2 collects the example machine readable instructions of engine.

Figure 10 is the example visitor table that tuning minute of paid close attention to demographic example visitor and exposure minute are shown.

Figure 11 is the schematic diagram collecting engine according to the example visitor of instruction structure of the present disclosure.

Figure 12 be for determine average visitor's parameter by being used for collecting multiple visitor, comprise example pay close attention to demographics and example pay close attention to classification exemplary unit parameter calculate.

Figure 13 is that the example independent parameter for determining the average visitor's parameter by being used for collecting multiple visitor calculates.

Figure 14 is the example probable value and the cumulative probability that are collected engine generation by the example visitor of Fig. 1 and Figure 11.

Figure 15 represents the process flow diagram that the example visitor that can be performed to realize Fig. 1 and Figure 11 collects the example machine readable instructions of engine.

Figure 16 is the schematic diagram of the tuning engine of example context according to instruction structure of the present disclosure.

Figure 17 to Figure 19 is the process flow diagram that representative can be performed to realize the example machine readable instructions of the tuning engine of example context of Fig. 1, Figure 10 and Figure 16.

Figure 20 is example record (crediting) figure of the example categories that the viewing of collecting minute is shown.

Figure 21 is the schematic diagram of the example ON/OFF detecting and alarm according to instruction structure of the present disclosure.

Figure 22 is the process flow diagram that representative can be performed to realize the example machine readable instructions of the example ON/OFF detecting and alarm of Fig. 1 and Figure 21.

Figure 23 be can perform Fig. 6 to Fig. 9, the instruction of Figure 15, Figure 17 to Figure 19 and/or Figure 22 collects the example processor platform of engine and example ON/OFF detecting and alarm schematic diagram with the tuning engine of the example context realizing Fig. 1, Fig. 2, Figure 10, Figure 16 and/or Figure 21, example.

Embodiment

What market survey personnel sought is understands spectators' composition and media (such as, radio programming, TV programme and/or internet media) scale, make to set up and to expose with spectators and demographics forms (being collectively referred to as " spectators are formed ") corresponding advertising rates herein.As used herein, " media " refer to content and/or the advertisement of any kind that is that presented by information presentation device (such as, TV, radio, computing machine, smart phone or flat board) or that can be presented by information presentation device.In order to determine that various aspects that spectators form (such as, which kinsfolk current is watching the demographics of the specific part of media and the correspondence of this kinsfolk), market survey personnel are by recruiting any amount of client as group member to perform audience measurement.These group members be recruited will be monitored audience membership (kinsfolk), they disclose and/or share their media exposure custom and consensus data, to contribute to carrying out market research.Audience measurement entity monitors media exposure custom (such as, watch, listen to) of the audience membership recruited usually by audience measurement system (such as, measurement mechanism and people meter).Audience measurement is usually directed to the identity of the media determined in the upper display of media presentation device (such as, TV).

Some audience measurement systems by physics mode be connected to media presentation device (such as, TV) with by gathering channel number, audio signature and/or (directly or indirectly) identify that the joint destination code just shown identifies current tune is which channel.The voice-frequency cable that can input by the audio frequency output of media presentation device being connected to audience measurement system, realizes the physical connection between media presentation device and audience measurement system.In addition, audience measurement system prompting and/or accept the input of audience membership, exposes to disclose which kinsfolk current the media that media presentation device presents.

As mentioned above, audience measurement entity can adopt audience measurement system, and these audience measurement systems comprise the device of such as people meter (PM), has the one group of input (such as, load button) being assigned to corresponding kinsfolk separately.PM is a kind of electronic installation, to be usually arranged in media exposure (such as, the watch) region of monitored family and one or more in contiguous audience membership.PM indicates them to be present in media exposure region (living room such as, residing for televisor) information gathered about family spectators by prompting audience membership by the input key tasking them that divides on (such as) pressing PM.When family's member selection during the input of their correspondences, PM identifies which kinsfolk that exist is, comprises other demographic information (such as, name, sex, age, receipts etc.) associated with kinsfolk.But if there is visitor in family, PM comprises at least one input (such as, load button) selected for visitor.When have selected visitor's load button, PM points out visitor to input age and sex (such as, by keyboard, by the interface etc. on PM).

PM can work in coordination with basic measurement device (such as, basic measurement instrument) and measure one or more signal associated with media presentation device together.Such as, basic measurement instrument can surveillance television machine, and with determination operation state, (such as, TV is energising or power-off, media apparatus power sensor) and/or identify the media (such as, identifying the program that televisor is just presenting) that media apparatus shows and/or issues.PM and basic measurement instrument can be devices separately and/or can be integrated in individual unit.Basic measurement instrument is by monitoring the audio frequency that monitored media presentation device exports and/or video as described above via cable and/or wirelessly gather audience measurement data.The audience measurement data of basic measurement instrument collection can comprise tuning information, signature, code (such as, embed in broadcast medium or otherwise broadcast with broadcast medium) and/or be exposed to the quantity of corresponding kinsfolk and/or the identity of the media that media presentation device (such as, TV) exports.

The data that PM and/or basic measurement instrument are collected can be stored in memory and via one or more network (such as, internet) be transferred to by market survey entity (such as, Nelson's market survey company limited of the U.S. (The Nielsen Company (US), LLC)) data back that manages.Usually, this data are together with the data aggregate collected from a large amount of PM and/or basic measurement instrument that monitor a large amount of group member family.The data of this collection and/or polymerization can be processed further, with determine one or more pay close attention to the statistics associated with family behavior in geographic area.Family's behavioral statistics can include, but is not limited to home media device be transferred to certain station minute quantity, family group member and/or one or more visitor use (such as, watching) home media device minute quantity, spectators demographics (can reflect with statistical based on group member's data) and media apparatus opens or closes time situation.Although example described herein have employed term " minute " (such as, " tuning minute of family ", " exposure minute " etc.), other time measurement of paying close attention to any can be adopted, not restriction.

Correctly be arranged in group member family in order to ensure audience measurement system, on-site service personnel visits each group member family as usual, home media assembly is assessed, physically install (such as, connecting) PM and/or basic measurement instrument to be to monitor the media presentation device of family (such as, televisor), and train kinsfolk how with PM alternately to gather accurate audience information.If one or more aspect that PM and/or basic measurement instrument are installed not inadvertently is interrupted (such as, connect be disconnected from media apparatus to the audio cable of basic measurement instrument), then on-site service personnel's visit is subsequently necessary.In order to allow to use collected family data (such as, to meet the mode of the statistical sample size of acceptance) in a reliable fashion, need PM and/or the basic measurement instrument of relatively large amount.Each this PM and/or basic measurement instrument all relate to one or more installment work and installation cost.Like this, the effort for the increase statistic validity (such as, by increasing group member's size and/or diversity) of paid close attention to colony causes the cost utilizing PM and/or basic measurement instrument to realize group member family correspondingly to increase.

In order to increase the sample size of family's behavioral data and/or reduce to be configured relevant cost to utilizing PM and/or basic measurement instrument to group member family, illustrative methods disclosed herein, device, system and/or goods adopt media measuring instrument (MM) to collect family group member behavioral data.Exemplary MM disclosed herein is different from conventional P M and/or the basic measurement instrument of the physical connection being included in media presentation device (such as, televisor).In example disclosed herein, MM gathers audio frequency and without the need to the physical connection of media apparatus.In some examples, MM comprises one or more microphone, to gather environment audio frequency in the room shared by media apparatus.In some this example, MM gathers the code of one or more entity embedding (such as, final distributor's Audiocode (FDAC)), and do not comprise selected by one or more family group member to identify current one or more input watching media apparatus of which group member.Which kinsfolk one or more model of illustrative methods disclosed herein, equipment, system and/or product application collects and is exposed to concrete media program to collect MM data, instead of collects directly from spectators' composition data of group member.This exemplary technology that collects will hereafter more specifically describe, and be called as herein " personnel collect ".In addition, one or more model of illustrative methods disclosed herein, device and product application collect the correspondence of visitor in each family and these visitors age/population characteristic.In other words, example disclosed herein facilitates the mode at visitor's age of the probability determining family's exposure behavior, some visitors and/or correspondence with the random fashion of the cost avoided the additional PM device in group member family and install.

In some examples, a family comprises two or more media apparatus, such as, be positioned at first televisor in the first room and be positioned at second televisor in the second room.If group member family comprises the first measuring instrument and the second measuring instrument that are physically connected to the first televisor and the second televisor, even if this audio frequency then from the first televisor propagates into second room (and/or vice versa) with the second televisor, physical connection also identifies which voice data comes from which televisor in family clearly.The situation that the media play in a room can be heard another room (it also can have media presentation device and adjoint measuring instrument) and/or detect is called as " spilling " (spillover) herein.If group member family comprises the MM laying respectively at the first room and the 2nd MM being positioned at the second room, then " (the detecting) heard " can be thought by the 2nd MM the media (and/or vice versa) being present in the second room mistakenly from the spilling voice data in the first room.Given the credit to by MM and appear at a room but the media tuning event in fact appearing at different the second room (such as, owing to overflowing) is called as herein " environment is tuning ".In other words, because MM comprises microphone to collect the audio frequency sent from media apparatus, so there is such possibility, that is, obtain at a MM in the first room and/or detect the audio frequency of the media apparatus from vicinity (such as, second) room.Environment is tuning to be with the difference of " truly tuning ", truly tuningly to appear at when the media exposure of the actual media be presented on media presentation device (such as, televisor) that the room that is positioned at this MM associates correctly is given the credit to this media presentation device by MM.Illustrative methods disclosed herein, device, system and/or product application model identify the example with true (rationally) tuning environment distinguished tuning (such as, owing to overflowing).Similarly, illustrative methods disclosed herein, device, system and/or product application model identify the example when media presentation device is opened distinguished with the example when media apparatus is closed.This is for being avoided when not occurring believing that this exposure is important when being exposed by the media.Such as, if kinsfolk is in the first room of closed condition at relevant media presentation device, but the relevant measuring instrument in this first room detects the audio frequency from the second media apparatus in the second room, then this event recognition is overflow and do not believe that this detection is actual media exposure by example disclosed herein.

Forward Fig. 1 to, exemplary media distributional environment 100 comprises the network 102 (such as, the Internet) of the group member family that can be connected to communicatedly in interest region (such as, goals research landform 104).In the example illustrated in Fig. 1, some group member family 106 comprises people meter (PM) and media measuring instrument (MM) 106, and some other group member family 108 only comprises MM to gather home media exposure information.The family with MM and PM is called as MMPM family 106 herein.The family not having PM still to have MM is called as MMH (media measuring instrument family) 108 herein.The behavioural information of being collected by exemplary MMPM 106 and exemplary MMH 108 via exemplary network 102 be sent to exemplaryly collect engine 110, exemplary guest collects engine, the tuning engine of exemplary environments 120 and/or exemplary ON/OFF detecting and alarm 130, to analyze.As mentioned above, because MMH 108 does not comprise PM, so they do not comprise selected by kinsfolk, in order to identify the current physical button input watching specific media of which kinsfolk, and they do not comprise selected by family visitor, in order to identify the physical button input at age and/or gender information.Therefore, illustrative methods disclosed herein, device, system and/or goods carry out modeling to Family characteristics, and these Family characteristics predict that concrete kinsfolk is watching the possibility of just accessed in MMH 108 media identified.

The example home comprising PM collects group member's attendance data.As used herein, " group member's attendance data " comprising: (a) media identification data (code such as, being embedded in the code in media, signature, channel tuning data etc. or sending by media, signature, channel tuning data etc.); And (b) personal information, it identifies the kinsfolk of correspondence and/or the visitor that currently watch/check/listen and/or access the media identified.On the other hand, MMH family 108 only comprises MM to collect media data.As used herein, " media data " and/or " media identification information " is used interchangeably io, and represent that the information relevant to media identification is (such as, code, signature etc.), but do not comprise to currently watching/check/listen and/or access the personal information that the kinsfolk of the media identified and/or visitor identify.As described in further detail below, personal identification data (PID) collects to from the media data collected by MMH family 108 by illustrative methods disclosed herein, device, system and/or goods.

Although example disclosed herein relate to code reader and collect code, technology disclosed herein also can be applied to collect signature and/or channel tuning data to identify the system of media.Audio frequency watermark is the technology for identifying the media that such as television broadcasting, radio broadcasting, advertisement (TV and/or radio), the media downloaded, Streaming Media, the media that encapsulate in advance etc. are such.Existing audio watermarking technique passes through by one or more Audiocode (such as, one or more watermark) (such as, media identification information and/or can map to the identifier of media identification information) embed audio frequency and/or video component to identify media.In some examples, the audio or video component with the signal characteristic being enough to hiding watermark is selected.As used herein, term " code " or " watermark " are used interchangeably, and be defined as representing the object that identifies media or in order to such as tuning other object like this (such as, bag identify header), can with the audio or video of media (such as, program or advertisement) send or insert or any identifying information (such as, identifier) of being embedded in the audio or video of media.As used herein, " media " refer to audio frequency and/or vision (static or motion) content and/or advertisement.In order to discriminating watermark media, watermark is extracted and for access map to the table of the reference watermark of media identification information.

Different from the media monitoring technology based on the code of media and/or watermark that are included in or are embedded in monitoring, usually the inherent feature of one or more media of monitoring is used in monitoring time interim, to generate the fully unique agency for media based on the media monitoring technology of fingerprint or signature.This agency is called as signature or fingerprint, and any aspect of presentation medium signal can be taked (such as, form audio frequency and/or the vision signal of monitored media representation) any form (such as, a series of digital value, waveform etc.).Good signature is such signature, that is, be repeatably when processing same media representation, but be unique relative to other (such as, different) expressions of other (such as, different) media.Therefore, term " fingerprint " and " signature " can exchange herein, and are defined as herein representing for identifying the media produced from one or more inherent feature of media.

Media monitoring based on signature is usually directed to determine (such as, generate and/or collect) represent the media signal that exported by monitored media apparatus (such as, sound signal and/or vision signal) signature and by monitored signature with correspond to one or more reference signature of known (such as, reference) source of media and compare.Various standard of comparison (such as, cross correlation value, Hamming distance etc.) can be assessed, to determine whether monitored signature matches with specific reference signature.When finding the coupling between monitored signature and a reference signature, the media of monitoring can be identified as corresponding to the specific reference medium represented by the reference signature of the signature of the monitoring with coupling.Because such attributes such as the identifier of such as media, presentative time, broadcasting channel are collected for reference signature, so these attributes can then associate with monitoring media (its monitored signatures match reference signature).For identifying that the example system of media is known for a long time based on code and/or signature, and the openest in the United States Patent (USP) 5,481,294 of Thomas, by its this by reference entirety be incorporated to.

Personnel collect

Fig. 2 is the schematic diagram collecting the exemplary realization of engine 110 of Fig. 1.In the example exemplified by Fig. 2, collect engine 110 and comprise visitor and collect engine 112, people meter (PM) interface 202, media measuring instrument (MM) interface 204, sorter 206, weighting engine 210 and probabilistic engine 212.As specifically described further below, exemplary guest collects engine 112 and adopts exemplary one or more part collecting engine 110.The example probability engine 212 of Fig. 2 comprises Exemplary metrology manager 214, exemplary cell probabilistic engine 216 and exemplary independent distribution engine 218.The exemplary cell probabilistic engine 216 of Fig. 2 comprises example classes and is applicable to manager 220, exemplary minute polymerizer 222 and exemplaryly collects engine 224.The exemplary independent distribution engine 218 of Fig. 2 comprises example categories delimiter 226, exemplary ratios manager 228 and exemplary distribution engine 230.

In operation, exemplary PM interface 202 obtains people meter's data from any PM exemplary group member family 104 and whole PM.Particularly, exemplary PM interface 202 obtains the PM data from the PM being positioned at exemplary MMPM family 106 (that is, having the family of MM device and PM device).PM device has input (such as, selecting to identify that they appear at the button be exposed in the spectators of media separately for each kinsfolk).In some examples, MMPM family 106 and the specific geographic area paid close attention to are (such as, in the whole nation (being sometimes referred to as " national people meter " (NPM))) relevant, and in other example, the subset of MMPM family 106 and the specific geographical area paid close attention to (such as, domestic city (such as, Chicago)) relevant, and sometimes referred to as " National Officer's measuring instrument " (LPM).

Such as, if expect to analyze the Xia Luote market area (DMA) of specifying, then exemplary PM interface 202 gathers the data from the LPM family in the time zone (such as, Eastern Time Zone) corresponding with the DMA expected.In some examples, the data expected can flow back to one or more storage repository, and exemplary engine 110, the tuning engine of exemplary environments 120 and/or the exemplary ON/OFF detecting and alarm 130 of collecting can from this one or more storage repository retrieve data.The PM data (group member's attendance data) from (having PM's and MM) group member family 104 are collected, obtain and/or gathered to the exemplary PM interface 202 of illustrated example, and for everyone record or the aggregated media exposure minute as one or more possible audience membership (such as, viewer) of corresponding media in family.In other words, the group member's attendance data gathered is in individual rank instead of family's rank, this facilitates the ability generating individual probability, as below will be specifically described further.

The example classes device 206 of Fig. 2 is classified to obtained group member's attendance data with many classifications, such as by the age, by sex, by whether be scale the be family of (such as, single family) or scale be two or larger family's (such as, have in family two or more individual), by platform/alliance, by kind and/or pass through the period.In some examples, classification comprises and race, blood lineage, geography, language, type that city is relevant to small towns etc.In other example, classification comprises the existence of age of householder, room location (such as, living room, principal bedroom, all the other bedrooms etc.) and/or child.If one or more classification improves result, then it can be used to analyze, and can remove the classification not demonstrating and improve or cause negative influence in analysis simultaneously.

Classification used herein refers to and the classification that the exposure of collecting minute (also referred to as " watch minute ") is associated.Period that classification can include but not limited to be associated with the exposure minute of collecting (5 o'clock to the 6 o'clock morning of such as the week, the 1:00 AM etc. at 10 in evening to next day on Sunday), the platform (such as WISN, WBBM etc.) be associated with the exposure minute of collecting, be associated with the exposure minute of collecting age/gender (such as 2-5 year the male sex, 35-44 year women etc.) and the type (such as children's programs, household keep in repair program, music program, I sports program etc.) that is associated with the exposure minute of collection.In other example, sorter 206 according to education degree (such as 8 years old or following, 9 years old to graduating from the high school, junior college academic or more etc. to undergraduate course educational background, master), life stage (such as do not get married, young families, ripe family, aged family, retirement etc.) and/or the quantity of media playing apparatus (televisor in such as family) obtained group member's attendance data is classified.Based on the change between the data available of such as one or more age/gender grade, in a different manner one or more combination of platform/alliance/type/ascribed characteristics of population can be classified.Such as, some local markets have 10 platforms, and wherein the sample size of the 45-54 year male sex can embody the data sample size in the statistical significance of 7 platforms in 10 platforms.In other example, local market can have less platform, and wherein the size of age/gender grade is enough to support statistics.In the example that some are such, regulate age/gender grouping (such as from 40-45 year the male sex be adjusted to 40-50 year the male sex) increase usable samples size, to realize required statistical significance.

In order to by group member's attendance data (such as exposure minute, usually also referred to as " viewing minute ") collect media data, example PM interface 202 identifies National Officer's measuring instrument (LPM) data of having collected in threshold time period.Under relative scale, when to when such as tv exposure processes, can calculate exposure index (represent the exposure of LPM purpose data classifying minute levels of precision) according to the mode meeting equation (1):

equation (1)

In the example of shown equation (1), the actual LPM that the LPM collected calculating each paid close attention to classification watches the number of minutes and each paid close attention to classification watches the ratio of the number of minutes, as exposure index.

Can manually, automatically, regularly, irregularly and/or according to the example exposure index arranging calculation equation (1), thus empirically verify success and/or the precision of the disclosed herein behavior that collects.Compared with the exponential quantity departing from (1), represent higher precision close to the exponential quantity of (1).According to the type of the classification be associated with collected exposure minute, corresponding exposure index value can realize higher or lower precision based on the time of collected data.Fig. 3 is the exemplary plot 300 by the exposure index value of period.In the example depicted in fig. 3, Figure 30 0 comprises the x-axis of time period value 302 and the y-axis of corresponding exposure index value 304.The exponential quantity data point being labeled as " 1 week " is rendered as generally closer to exponential quantity 1.00, and the exponential quantity data point being labeled as " 3 weeks " is rendered as generally away from exponential quantity 1.00.In other words, the group member's attendance data collected in the recent period more makes exponential quantity close to 1.00, thus the group member's attendance data embodying collection before than 1 week higher collect precision.

As mentioned above, the data volume collected in the recent period reveal the data of relatively early collecting higher collect precision.However, some data comparatively are early still useful, but when reflecting lower accuracy, this weighted ratio Recent data compared with early data is little.Exemplary weights engine 210 Applicative time weight, and corresponding weighted value is applied according to number of days from Data Collection.Larger weighted value is applied to comparatively near data of collecting.In some instances, the weighted value of the exposure that is applied to tuning minute of collection and collect minute is the ratio based on timestamp associated with it.Such as, when collection part was more Zao than recent minute for the timestamp be associated, weight less for ratio is applied to minute (such as tuning minute, exposure minute) that a part is collected.

Fig. 4 shows the exemplary weights allocation table 400 that exemplary weights engine 210 generates and/or configures.In the example depicted in fig. 4, minute (namely MMPM family 106 obtains exposure via PM device (" A " OK), personalized group member's attendance data), and obtain tuning minute of family (that is, tuning in the family and do not carry out personalized minute for the specific people in family) via MM device (" B " OK).Example personalized spectators group member and tuning minute of family is collected within the period in seven (7) skies.In this way, the nearest date (when the day before yesterday 402) has been associated with than the large weight of any personalized spectators group member since previous dates and/or tuning minute of family.According to the category combinations of required given family, can Further Division " A " row example personalized group member minute.As mentioned above, describe the classification of Family characteristics and can comprise given age/sex, household size, the platform of viewing, period, TV quantity, life stage, education degree and/or other ascribed characteristics of population.In order to be described, in following example, the age/gender classification of family is the 45-54 year male sex, and the pay channel (type) in the period between tuning platform and the 6 pm to 7 o'clock of the week is associated.

In the example depicted in fig. 4, unified weighted value is applied to personalized group member minute and family the first six (6) sky of tuning minute by weighting engine 210, and weighted value six (6) is applied to the most current date.Although other value discloses value six (6) like that as used herein above, being using only exemplary purposes and not being construed as limiting of this value.In operation, the exemplary weights engine 210 of Fig. 2 can adopt any weighted value, wherein when the value of the day before yesterday to be greater than before the day before yesterday one day or more the value in sky.Exemplary weights engine 210 can generate the weighted sum of collected personalized spectators' group member exposure minute (hereinafter referred to as " exposure minute ") according to the mode meeting example equation (2), and can generate the collected family weighted sum of tuning minute according to the mode meeting example equation (3).

equation (2)

equation (3)

In the example that equation (2) and equation (3) illustrate, W ₁represent and be less than W ₂weighted value, W ₂be and work as the weighted value exposing minute value the day before yesterday and be associated.In addition, d represents a date in n the date of the data of collecting before the day before yesterday, EM _drepresent the exposure minute of the corresponding day before the day before yesterday, TM _dtuning minute of the family of the corresponding day of expression before the day before yesterday, EM _crepresent the exposure minute when the day before yesterday, TM _crepresent tuning minute of the family when the day before yesterday.

For the sample data shown in the example shown in Fig. 4, (such as exposure minute is respectively the day 1-6 of 20,10,10,0,0,10, exposure minute be 40 work as the day before yesterday, family within tuning minute, be respectively 40,30,50,0,0,30 day 1-6 and family within tuning minute, be 50 work as the day before yesterday), application example equation (2) obtains weighting exposure minute value 290, and application example equation (3) obtains weighting family tuning minute value 450.In some instances, probabilistic engine 212 calculate have above-mentioned paid close attention to category combinations (period between such as on Monday to the 6 pm of Friday by 7 o'clock be tuned to pay channel 45-54 year the male sex) MM group member's (such as only have MM device and do not associate the group member family of PM device) actual watch this tuning session collect probability.Example probabilistic engine 212 collects probability by weighting exposure minute (such as 290 minutes) being calculated divided by weighting family tuning minute (such as 450 minutes), show that the chance that the MM group member with same home category combinations is associated with this tuning behavior is 64.4%.Although example disclosed herein is probability calculation, in some instances, probability may be calculated as value result be restricted between 0 and 1.Such as, probability can be calculated as the ratio of probable value divided by (1-probability).If necessary, probability can be transformed back the form of probability.

But, although Market Researcher can adopt and specific pay close attention to category combinations, but when applying different probability calculation technology, can according to the probable value precision sharing usable samples size that specific institute pays close attention to the correspondence of the family of category combinations and improve correspondence.As hereafter more specifically described, if collect with paid close attention to category combinations (such as in the afternoon 6 o'clock to 7 o'clock be tuned to pay channel 45-54 year the male sex, there are three kinsfolks and a televisor, and householder has university's credit or bachelor's degree) the LPM data that are associated are greater than threshold value, then and orthant probabilities technology can obtain the probable value with acceptable precision.The precision that accepts used herein relates to the sample size can and/or being required to set up the result with statistical significance.But when collected National Officer's measuring instrument (LPM) data be associated with paid close attention to category combinations break threshold value by a fall, orthant probabilities technology can only obtain unacceptable low probable value precision.On the contrary, when the collected LPM data be associated with paid close attention to category combinations are lower than (such as lower than contributing to the threshold value that one or more calculates the result to obtain having statistical significance) during threshold value, exemplary method disclosed herein, equipment, system and/or goods adopt independent distribution probability calculation.

Example classification manager 214 pairs of classifications of Fig. 2 and/or pay close attention to category combinations and identify, and determine specific to pay close attention to the family whether category combinations has number of thresholds in donor pond.As implied above, donor pond can be local place (National Officer's measuring instrument (LPM), such as, group member family in paid close attention to geographic area 104).But along with the size of paid close attention to geographic area diminishes, the respective amount of the qualified family of mating with paid close attention to category combinations also reduces.In some cases, the quantity of qualified family, lower than threshold value, makes one or more of conventional probability computing method (such as orthant probabilities) present poor predictive ability and/or cannot produce the result of statistical significance.On the other hand, when the donor pond of family exceedes threshold value, these conventional probability computing method (such as orthant probabilities) present satisfied predictive ability under industrial standard.

In operation, the example classification manager 214 of Fig. 2 pays close attention to the condition test of classification formation logic "AND" for one group of institute.Such as, if pay close attention to classification and comprise (1) specific platform, (2) specific time period, (3) kinsfolk of specific quantity, (4) given age, (5) particular sex, (6) televisor of specific quantity in family, (7) the specific level of education of householder, and (8) specific life stage, then class manager 214 determine whole eight kinds pay close attention to classification combination whether represented by the family of the number of thresholds in donor pond.If then example classification manager 214 calls exemplary unit probabilistic engine 216, to calculate the probable value of the category combinations occurred in MMH family 108.In general, when family's quantity of shared paid close attention to category combinations ((1) such as-(8) item) is greater than threshold value, the confidence level of the correspondence of the probability calculation adopting orthant probabilities technology to carry out is gratifying.

When Market Researcher seeks the probabilistic information of the 50 years old male sex watching pay channel in the afternoon between 6 to 6:30, the example categories of illustrated example is applicable to manager 220 and identifies that in the classification group of the prior foundation existed, which is best suited for required task.In other words, concrete and/or other unique research of Market Researcher requires accurately to coincide with the existing classification group of being collected by LPM and/or NPM device.On the contrary, example classification is applicable to immediate category combinations that manager 220 identifies industrial standard and/or other expected data and is included in the 45-54 year male sex between 6 pm to 7 o'clock.The example minute polymerizer 222 of illustrated example identifies the family's tuning minute sum in all families be associated with identified immediate category combinations, and identify with in these families 45-54 year the male sex exposure that is associated minute total.Such as, minute polymerizer 222 can identify the qualified family (also can comprise the people outside the 45-54 year male sex in such as family) of 45 (45) the individual 45-54 of the comprising year male sex, in these qualified families, watch the platform of paying in the afternoon between 6 o'clock to 7 o'clock, three kinsfolks have a televisor and householder has university's credit or bachelor's degree.

In these 45 (45) individual qualified families, tuning minute polymerizer 222 can identify tuning minute of 200 (200) family altogether, but wherein only having 102 (102) minutes with the 45-54 year male sex is associated.The example of illustrated example collects engine 224 according to the mode meeting example equation (4) to calculate the exposure minute of the 45-54 year male sex and the total family ratio of tuning minute for all qualified families, as collecting probability.

equation (4)

In the illustrated example of equation (4), adopt above-mentioned open example to collect probability be 0.51 (namely in this example, 102 exposures minute divided by 200 tuning minutes).In some instances, the probable value that calculates of exemplary unit probabilistic engine 216 is retained based on normal distribution (probable value such as calculated and random number or pseudo random number compare) and/or is collected in MMH family 108.When the probable value calculated is greater than random number, then think that the kinsfolk with paid close attention to category combinations is at viewing tuning part.In other words, the exposure data of family's tuning data as paid close attention to category combinations is collected in MMH family 108.On the other hand, when the probable value calculated is less than random number or pseudo random number, then think that the kinsfolk with paid close attention to category combinations does not watch tuning part.In other words, family's tuning data can not be collected in MMH family 108.

As discussed above, when combination by all the paid close attention to classifications of the family's representative being less than number of thresholds in donor pond, orthant probabilities numerical procedure may not show the confidence level being considered suitable for statistical research.As a rule, pay close attention to study geography in may be relative to high with the quantity of the family of single paid close attention to categorical match.But, because with the addition of additional institute pays close attention to classification, so there is the quantity minimizing of the family that inclusive is mated for all such classifications.In some environment, the quantity of coupling family available in donor pond after performing logical "and" to all paid close attention to classifications finally causes population lower than the donor pond of threshold value, and this may not represent statistical confidence when applying said units probabilistic technique.In such an example, traditional orthant probabilities technique computes that prevents from probabilistic engine 212 adopting whether the probability of the exposure behavior for paid close attention to category combinations should be given the credit to paid close attention to family (such as, whether the exposure of gathered family (tuning) behavior should be given the credit to family 45-54 year the male sex).And when have expectation pay close attention to category combinations family's quantity lower than threshold value time, example probability engine 212 calls exemplary independent distribution engine 218.As will be described in further detail below, replace using and the family pond of all paid close attention to categorical match, adopt use when watching probability with calculating pay close attention in classification some pay close attention to the family that classification matches.

In operation, the example categories delimiter 226 of Fig. 2 is to having the key forecast factor of identical group (namely, one group pay close attention to particular category in classification) donor pond in all families of (such as, LPM collects in geography, such as Xia Luote DMA) identify.In some instances, the reflection of the key forecast factor shows one group of classification of relatively large Degree of Success than other category combinations.Such as, first group of key forecast factor can comprise to pay close attention to geographical relevant first group of classification (such as, the geographically sunscreen product of contiguous coastal resorts, or be geographically close to the skiing product on mountain range).Although example disclosed herein represents National Officer's measuring instrument (LPM), such example is not limited thereto.In some instances, the collection that national people meter (NPM) can be used as the relatively large region (as country) of reflection is geographical.Particularly, the subgroup of exemplary eight (8) original paid close attention to classifications can comprise: the family of family of the family of (1) and household size categorical match, (2) and identical member's sex categorical match and (3) and identical member age's categorical match.In other words, even if original 8 exemplary paid close attention to classifications comprise 3 classifications above-mentioned, when identifying family from data available pond, consider to remove all the other classifications.Such as, remove and following all the other relevant classifications: the family of the family of the family of the family of the family of (4) and identical tuning platform categorical match, (5) and identical education categorical match, (6) and the televisor categorical match of equal number, (7) and identical period categorical match, (8) and identical life stage/household size categorical match.

In the illustrated example, because donor pond is only made up of MMPM family 106, so example categories delimiter 226 is retrieved for the available Home of satisfied institute concern scale/sex/age criterion and/or obtains total family tuning minute value and total exposure minute value (such as, from above tolerance (1), (2) and (3)).Such as, if institute's concern scale/sex/age criterion is the household size for two or more individuals with men age 45-54, then example categories delimiter 226 is from described scale/multiple family of sex/age sub-centralised identity.

Fig. 5 illustrates the example categories subset created by the independent distribution engine 226 in the example of Fig. 2 and maps 500.Exemplary independent distribution engine collects family tuning minute and exposure minute from the subset of paid close attention to classification.In the illustrative example of Fig. 5, map 500 and comprise the total family relevant to the key forecast factor classification 502 of scale/age/gender and within tuning minute, to count and total exposure minute counts.In this example, classification delimiter 226 identifies 200 (200) families altogether of mating with scale/sex/age criterion.200 families comprise altogether that 4500 tuning minutes (namely, identify tuning platform but do not identify corresponding kinsfolk minute) and 3600 exposures minute altogether (such as, for the platform that identifies and also for the individual of mark that exists in spectators minute).

One or more all the other classifications of paying close attention to fallen into outside key forecast factor classification selected by the exemplary ratios manager 228 of Fig. 2, to determine corresponding available coupling family, tuning minute of family and exposure minute.All the other classifications exemplary can be called as the secondary predictor of the probability that impact is exposed by the media or secondary classification.And example key predictor classification disclosed herein comprises household size, sex and age, illustrative methods, equipment, system and/or goods can comprise for the key forecast factor any other, the classification of type that is additional and/or that substitute.In addition, although exemplary secondary classification disclosed herein comprises tuning platform, education, the quantity of media presentation device (televisor), period and life-span, illustrative methods, equipment, system and/or goods in addition and/or alternatively can comprise the classification of other type any as secondary classification.

Such as, the Commercial banks device 228 in illustrative example selects one or more secondary classification to determine the coupling family of respective numbers, tuning minute of family and exposure minute.Moreover, as mentioned above, in order to discuss convenience when illustrative methods disclosed herein, equipment, system and/or goods, there is employed herein chronomere " minute ", make it possible to consider ad lib one or more additional and/or replaceable chronomere (such as, second, sky, hour, week etc.).In the illustrative example of Fig. 5, Commercial banks device 228 identifies tuning post classification 504 (such as, secondary a paid close attention to classification) have 80 (80) individual with expect pay close attention to platform (such as, platform " WAAA ") family of mating, wherein these families have collected tuning minute of 1800 families and 1320 exposures minute.In addition, the exemplary ratios manager 228 of Fig. 2 selects education classification 506 (such as, one in secondary paid close attention to classification), and determine that 110 (110) individual families mate (such as with paid close attention to expectation level of education, householder has education in 9 years to the family graduated from the high school), wherein, these families have collected tuning minute of 1755 families and 1200 exposures minute.In addition, the exemplary ratios manager 228 of Fig. 2 selects televisor quantity classification 508 (such as, secondary a paid close attention to classification), and determine that 100 (100) individual families mate with the televisor of the desired amt in family value, wherein, these families have collected tuning minute of 2100 families and 1950 exposures minute.Period classification 510 is comprised (such as by other example categories that the exemplary ratios manager 228 of Fig. 2 is considered, one in secondary paid close attention to classification), wherein, the period categorical match of 100 (100) individual families and expectation determined by the Commercial banks device 228 of Fig. 2, wherein, these families have collected tuning minute of 1365 families and 825 exposures minute.The exemplary ratios manager 228 of Fig. 2 also selects life stage/household size classification 512 (such as, one in secondary paid close attention to classification), and determine that 70 (70) individual families mate with the life stage/household size value of desired type, wherein, 1530 families are collected tuning minute and 1140 exposures minute by these families.

As a rule, the Commercial banks device 228 in illustrative example is tuning from the contingent family of this kind of family of only mating for the objective cross of all expectations with paid close attention to classification and expose and minute to identify tuning minute of family and contribute with the secondary classification to expose minute independently.After identifying each indivedual secondary classification contribution family's tuning minute value and exposure minute value, exemplary distribution engine 230 calculates the tuning tuning ratio of corresponding family and the exposure ratio with exposing minute value based on key forecast factor family.As will be described in further detail below, exemplary distribution engine 230 calculates the tuning ratio of family of being correlated with each paid close attention to secondary classification (such as, tuning post classification 504, educate classification 506, televisor quantity classification 508, period classification 510 and life stage/scale classification 512) and exposes ratio.In other words, example disclosed herein by calculating and/or the independent corresponding tuning ratio that additionally identifies for each secondary classification and independent corresponding exposure ratio, gather, calculate and/or identify one or more pay close attention to the contributing effect of secondary classification.As will be described in further detail below, the independent contributing effect of one or more secondary classification is aggregated tuning minute that calculates expectation and the exposure minute estimated.

In the illustrative example of Fig. 5, distribution engine 230 by tuning for the family relevant to tuning post classification 504 minute (such as, tuning minute of 1800 families) divided by tuning minute of total family relevant to key forecast factor classification 502 (such as, tuning minute of 4500 families), to calculate the tuning ratio 514 of corresponding tuning post classification.In addition, distribution engine 230 in illustrative example by the exposure relevant to tuning post classification 504 minute (such as, 1320 exposures minute) divided by the total exposure relevant to key forecast factor classification 502 minute (such as, the viewing minute of 3600 families), to calculate corresponding tuning post classification viewing ratio 516.For the purpose of illustrating, the tuning ratio 514 of the tuning post classification calculated is 0.4 (such as, tuning minute of 1800 families are divided by 4500 total exposures minute), and the tuning post classification calculated viewing ratio 516 be 0.37 (such as, 1320 exposures minute divided by 3600 total exposures minute).

The exemplary distribution engine 230 of Fig. 2 also calculates family's tuning ratio example and exposure ratio in conjunction with exemplary education classification 506.In the illustrative example of Fig. 5, distribution engine 230 by tuning minute of the relevant family of education classification 504 (such as, tuning minute of 1755 families) divided by tuning minute of total family relevant to key forecast factor classification 502 (such as, tuning minute of 4500 families), to calculate the tuning ratio 518 of corresponding education classification family.In addition, exemplary distribution engine 230 in illustrative example by the relevant exposure of education classification 506 minute (such as, 1200 exposures minute) divided by the total exposure relevant to key forecast factor classification 502 minute (such as, 3600 exposures minute), to calculate corresponding education classification exposure ratio 520.For the purpose of illustrating, the tuning ratio 518 in other front yard educational calculated is 0.39 (such as, tuning minute of 1755 families were divided by tuning minute of 4500 total family), and the education classification calculated exposure ratio 520 be 0.33 (such as, 1200 exposures minute divided by 3600 total exposures minute).

The exemplary distribution engine 230 of Fig. 2 also calculates family's tuning ratio example and exposure ratio in conjunction with example home televisor classification 508.In the illustrative example of Fig. 5, distribution engine 230 by tuning for the family relevant to home television set classification 508 minute (such as, tuning minute of 2100 families) divided by tuning minute of total family relevant to key forecast factor classification 502 (such as, tuning minute of 4500 families), to calculate the tuning ratio 522 of corresponding home television set classification family.In addition, exemplary distribution engine 230 in illustrative example by the exposure relevant to home television set classification 508 minute (such as, 1950 exposures minute) divided by the total exposure relevant to key forecast factor classification 502 minute (such as, 3600 exposures minute), to calculate corresponding home television set classification exposure ratio 524.For the purpose of illustrating, the tuning ratio 522 of home television set classification family calculated is 0.47 (such as, tuning minute of 2100 families were divided by tuning minute of 4500 total family), and the home television set classification calculated exposure ratio 524 be 0.54 (such as, 1950 exposures minute divided by 3600 total exposures minute).

The exemplary distribution engine 230 of Fig. 2 also calculates family's tuning ratio example and exposure ratio in conjunction with Exemplary periods classification 510.In the illustrative example of Fig. 5, distribution engine 230 by tuning for the family relevant to period classification 510 minute (such as, tuning minute of 1365 families) divided by tuning minute of total family relevant to key forecast factor classification 502 (such as, tuning minute of 4500 families), to calculate the tuning ratio 526 of corresponding period classification family.In addition, the exemplary distribution engine 230 of Fig. 2 by the exposure relevant to period classification 510 minute (such as, 825 exposures minute) divided by the total exposure relevant to key forecast factor classification 502 minute (such as, 3600 exposures minute), to calculate corresponding period classification exposure ratio 528.For the purpose of illustrating, the tuning ratio 526 of period classification family calculated is 0.30 (such as, tuning minute of 1365 families were divided by tuning minute of 4500 total family), and the period classification exposure ratio 528 calculated be 0.23 (such as, 825 exposures minute divided by 3600 total exposures minute).

The exemplary distribution engine 230 of Fig. 2 also calculates family's tuning ratio example and exposure ratio in conjunction with exemplary life stage/scale classification 512.In the illustrative example of Fig. 5, distribution engine 230 by tuning minute of the family that life stage/scale classification 512 is relevant (such as, tuning minute of 1530 families) divided by tuning minute of total family relevant to key forecast factor classification 502 (such as, tuning minute of 4500 families), to calculate the tuning ratio 530 of corresponding life stage/scale classification family.In addition, the example distribution engine 230 of Fig. 2 by the exposure relevant to life stage/scale classification 512 minute (such as, 1140 exposures minute) divided by the total exposure relevant to key forecast factor classification 502 minute (such as, 3600 exposures minute), to calculate corresponding life stage/scale classification exposure ratio 532.In this example, the tuning ratio 530 of the life stage calculated/scale classification family is 0.34 (such as, tuning minute of 1530 families were divided by tuning minute of 4500 total family), and the life stage calculated/scale classification exposure ratio 532 be 0.32 (such as, 1140 exposures minute divided by 3600 total exposures minute).

As mentioned above, each concern in the objective cross of classification there is the tuning ratio value of the family independently calculated and the independent exposure ratio value that calculates.The example distribution engine 230 of Fig. 2 calculates the tuning ratio value of all families (such as, the tuning tuning ratio of classification family of TV station 514, the tuning ratio in other front yard educational 518, the tuning ratio of home television set classification family 522, the tuning ratio of period classification family 526, the tuning ratio 530 of life stage/scale classification family) product, to determine the family of total expectation tuning minute 513.In addition, the example distribution engine 230 of Fig. 2 calculates the product of all families exposure ratio value (tuning TV station's classification exposure ratio 516, education classification exposure ratio 520, home television set classification exposure ratio 524, period classification exposure ratio 528 and life stage/scale classification exposure ratio 532), to determine the exposure minutes 536 of total expectation.Final independent distribution is calculated according to the mode consistent with example equation (5) by example distribution engine 230, and reflects the group member's behavior probability be associated with paid close attention to objective cross.

equation (5)

In example discussed above exposure and family in tuning minute, the independent distribution probability obtained is 0.52.In fact, the independent distribution probability obtained is interpreted as the male sex in age 45-54 year, this male sex lives in the family of three (3) individuals being classified as aged family, wherein householder be subject to nine (9) years education to graduating from the high school, have two (2) platform televisors in family, this male sex on Monday to Friday from the possibility of watching WAAA TV station during period of 9:00 to 12:00 in afternoon in the morning be 52%.

Although Fig. 2 to Fig. 5 is exemplified with the way of example collecting engine 110 realizing Fig. 1, one or more in Fig. 2 in illustrative element, process and/or device can be combined, divides, reconfigures, omits, removes and/or realize in any other way.In addition, although the visitor realizing Fig. 1, Fig. 2 and Figure 11 collects the way of example of engine 112 and will describe in further detail below, one or more in Figure 11 in illustrative element, process and/or device can be combined, divides, reconfigures, omits, removes and/or realize in any other way.In addition, although the way of example of the tuning engine 120 of the environment realizing Fig. 1 and example ON/OFF detecting and alarm 130 illustrates respectively and will describe in further detail below in Figure 10 and Figure 15, one or more in Figure 10 and Figure 15 in illustrative element, process and/or device can be combined, divides, reconfigures, omits, removes and/or realize in any other way.In addition, example people meter's interface 202 of Fig. 1, example classification device 206, exemplary weights engine 210, example media measuring instrument interface 204, example probabilistic engine 212, example categories manager 214, exemplary unit probabilistic engine 216, example categories is applicable to manager 220, example minute polymerizer 222, example collects engine 224, example independent distribution engine 218, example categories delimiter 226, example Commercial banks device 228, example distribution engine 230 and/or more usually example collect engine 110 and/or example visitor and collect engine 112 and can pass through hardware, software, firmware and/or hardware, the combination in any of software and/or firmware realizes.In addition, example average visitor's parameter (AVP) counter 1102 of Fig. 1 and Figure 16, example distribution engine 1104, example randomizer 1106, example visitor specifies device (assignor) 1108, example simultaneous tuning watch-dog 1602, example record (crediting) manager 1604, example platform comparer 1606, the tuning type of example specifies device 1608, example automatic gain controls watch-dog 1610, there is manager 1612 in code sample, example Modeling engine 1614, the tuning engine 120 of code sample stack-manager 1616 and/or more usually example context can pass through hardware, software, firmware and/or hardware, the combination in any of software and/or firmware realizes.Therefore, such as, example people meter's interface 202 of Fig. 1, example classification device 206, exemplary weights engine 210, example media measuring instrument interface 204, example probabilistic engine 212, example categories manager 214, exemplary unit probabilistic engine 216, example categories is applicable to manager 220, example minute polymerizer 222, example collects engine 224, example independent distribution engine 218, example categories delimiter 226, example Commercial banks device 228, example distribution engine 230, example average visitor's parameter (AVP) counter 1102, example distribution engine 1104, example randomizer 1106, example visitor specifies device 1108, example simultaneous tuning watch-dog 1602, example record manager 1604, example platform comparer 1606, the tuning type of example specifies device 1608, example automatic gain controls watch-dog 1610, there is manager 1612 in code sample, example Modeling engine 1614, code sample stack-manager 1616 and/or more usually example collect engine 110, example visitor collects engine 112, any one in the tuning engine 120 of example context and/or example ON/OFF detecting and alarm 130 can pass through one or more analog or digital circuit, logical circuit, programmable processor, special IC (ASIC), programmable logic device (PLD) (PLD) and/or field programmable logic device (FPLD) realize.

When any one reading in the equipment of this patent or system claims realizes to cover pure software and/or firmware, example people meter's interface 202 of Fig. 1, example classification device 206, exemplary weights engine 210, example media measuring instrument interface 204, example probabilistic engine 212, example categories manager 214, exemplary unit probabilistic engine 216, example categories is applicable to manager 220, example minute polymerizer 222, example collects engine 224, example independent distribution engine 218, example categories delimiter 226, example Commercial banks device 228, example distribution engine 230, example average visitor's parameter (AVP) counter 1102, example distribution engine 1104, example randomizer 1106, example visitor specifies device 1108, example simultaneous tuning watch-dog 1602, example record manager 1604, example platform comparer 1606, the tuning type of example specifies device 1608, example automatic gain controls watch-dog 1610, there is manager 1612 in code sample, example Modeling engine 1614, code sample stack-manager 1616 and/or more common example collect engine 110, example visitor collects engine 112, it is the such as storer comprising storing software and/or firmware that at least one in the tuning engine 120 of example context and/or example ON/OFF detecting and alarm 130 is clearly defined at this, digital universal disc (DVD), CD (CD), the tangible computer readable storage devices of Blu-ray disc etc. or memory disc etc.In addition, Fig. 1, Fig. 2, Figure 11, the example of Figure 16 and/or Figure 21 collects engine 110, example visitor collects engine 112, the tuning engine 120 of example context and/or example ON/OFF detecting and alarm 130 are except comprising Fig. 2, Figure 11, illustrative element in Figure 16 and/or Figure 21, also one or more element can be comprised outside process and/or device, process and/or device, or instead, Fig. 2, Figure 11, the illustrative element of Figure 16 and/or Figure 21, process and/or device can comprise illustrated element, process and device any or all in more than an element, process and/or device.

Illustrated in Fig. 6 to Fig. 9, Figure 15, Figure 17 to Figure 19 and Figure 22 collect engine 110 for the example realized in Fig. 1, Fig. 2, Figure 11, Figure 16 and Figure 21, example visitor collects engine 112, the process flow diagram of the representative example machine readable instructions of the tuning engine 120 of example context and example ON/OFF detecting and alarm 130.In these examples, machine readable instructions comprises the program performed by the processor of the processor 2312 shown in the example processor platform 2300 such as discussed below in conjunction with Figure 23.Program can realize with the software be stored on the tangible computer readable storage medium storing program for executing of the such as CD-ROM relevant to processor 2312, floppy disk, hard disk drive, digital universal disc (DVD), Blu-ray disc or storer, but the part of whole software and/or software can alternatively be performed by the device except processor 2312 and/or realize with firmware or specialized hardware.In addition, although describe example procedure with reference to illustrative process flow diagram in Fig. 6 to Fig. 9, Figure 15, Figure 17 to Figure 19 and Figure 22, alternatively can use that realization example collects engine 110, example visitor collects engine 112, other method multiple of the tuning engine of example context 120 and/or example ON/OFF detecting and alarm 130.Such as, the execution sequence of frame can change, and/or some frames in the multiple frames described can change, remove or merge.

As mentioned above, Fig. 6 to Fig. 9, Figure 15, the example process of Figure 17 to Figure 19 and Figure 22 can utilize such as hard disk drive, flash memory, ROM (read-only memory) (ROM), CD (CD), digital universal disc (DVD), high-speed cache, the tangible computer of random access memory (RAM) and/or machine readable media and/or wherein storing message reach any duration (such as, the time period extended, for good and all, momently, for adhoc buffer and/or for caching messages) other memory storage any or memory disc on the instruction of coding that stores (such as, computing machine and/or machine readable instructions) realize.As used herein, it is the computer readable storage means and/or the memory disc that comprise any type that term tangible computer readable storage medium storing program for executing is clearly defined, and gets rid of transmitting signal and get rid of transmission medium.As used herein, " tangible computer readable storage medium storing program for executing " and " tangible machine readable storage medium storing program for executing " is employed interchangeably.Addition or alternatively, Fig. 6 to Fig. 9, Figure 15, the example processor of Figure 17 to Figure 19 and Figure 22 can utilize such as hard disk drive, flash memory, ROM (read-only memory), CD, digital universal disc, high-speed cache, the non-transient computing machine of random access memory and/or machine readable media and/or wherein storing message reach any duration (such as, the time period of expansion, for good and all, momently, for adhoc buffer and/or for caching messages) other memory storage any or memory disc on the instruction of coding that stores (such as, computing machine and/or machine readable instructions) realize.As used herein, it is the computer readable storage means and/or the memory disc that comprise any type that the non-transient computer-readable medium of term is clearly defined, and gets rid of transmitting signal and get rid of transmission medium.As used herein, phrase " at least " for the Transitional Language in the preorder of claim, " to comprise " for open identical mode is open with word.

The program 600 of Fig. 6 starts at frame 602, and at frame 602, example people meter's interface 202 has the MM device PM data relevant to kinsfolk with the PM device acquisition of the example MMPM family 106 of both PM devices from being positioned at.As mentioned above, PM device has input (such as, for the button of each kinsfolk and visitor button to identify their respective existence current being exposed in the spectators of media).Example PM interface 202 identifies the data of the collection in the threshold time section on the same day, is weighted these data in order to the relative age according to described data.As mentioned above, in conjunction with example equation (1), when corresponding collection data are more recent, the accuracy of viewing index is better.Example classification device 206 based on one or more pay close attention to classification to obtain PM data classify (frame 604).In some instances, the particular home that sorter 206 one or more classification to the education degree of the televisor quantity of paying close attention in the particular home scale combining with such as paid close attention to age/gender, pay close attention to, the specific life stage paid close attention to, the TV station/alliance/type of specific viewing paid close attention to, the specific time period paid close attention to, family (family such as, with a televisor, the family with 2 to 3 televisors, have the family etc. of three or more platform televisors) and/or householder is relevant carries out classifying and/or identifying.Although at least one that the MMPM family 106 of relatively large amount will have in classification above-mentioned, in fact lesser amt MMPM family 106 by representative to carry out during market survey market survey pay close attention to all objective cross of classification.

As above in conjunction with Figure 4, exemplary weights engine 210 is according to the ratio application weight (frame 606) based on the number of days from the collection day of donor.Example media measuring instrument interface 204 also obtains family's tuning data (frame 608) from the media surveying instrument MMH family 108.According in the donor pond with all paid close attention to categorical match (such as, the donor pond of the MMPM family in interest region) in whether there is a number of thresholds family, example probabilistic engine 212 will call the corresponding probability calculation technology (frame 610) described in further detail below in conjunction with Fig. 7.

Fig. 7 comprises the additional detail of the illustrative example to Fig. 6.When produce probability time, the identification of example categories manager to use pay close attention to classification.As a rule, exemplary method disclosed herein, equipment, system and/or goods produce probability based on the objective cross of paid close attention to classification, such as, determine the possibility of watching: the male sex of (1) age 45-54 for following item; (2) live in the family of three people; (3) aged family is categorized as; (4) householder has nine (9) the years education to graduating from the high school; (5) family has two televisors; (6) WAAA TV station is watched; (7) viewing time is between the period of 9:00 to 12:00 in afternoon in the morning.Example categories manager 214 identify need watch (exposure) probability pay close attention to classification (frame 702), all exemplary seven classifications as mentioned above.Based on identify such as have the above example that the age is the 45-54 year male sex etc. the objective cross of concern classification, example categories manager 214 determine previously by the data available pond of exemplary weights engine 210 weighting whether comprise with pay close attention to classification all (such as, whole seven) objective cross number of thresholds family (frame 704) of mating.

Such as, assuming that the threshold value family quantity of mating with paid close attention to all categories is 30 (30), and data pool comprises the available Home (frame 704) of this threshold quantity, probabilistic engine 212 calls exemplary unit probabilistic engine 216 with via orthant probabilities technique computes probable value (frame 706).On the other hand, if data pool does not meet the threshold quantity of the family of 30 (such as, be less than 30 families) (frame 704), then example probabilistic engine 212 calls example independent distribution engine 218 with via independent distribution technique computes probable value (frame 708).

The way of example (frame 706) that Fig. 8 calculates exemplified with the orthant probabilities realizing Fig. 7.In the illustrative example of Fig. 8, classification is applicable to manager 220 and selects and/or limit tuning and viewing data, to be applicable to the classification (frame 802) previously set up.As mentioned above, when market researcher pays close attention to the male sex that the age is 50 years old, industrial standard group member data acquisition technology may not be the demographics being accurately applicable to expecting.On the contrary, can classify to industrial standard data available according to the male sex of the range of age between 45-54.Because expect to pay close attention to classification be the male sex for 50 years old age, identify will meet the maximally related classification grouping of market researcher so example categories is applicable to manager 220, in this example, most related category colony comprises the male population of age between 45-54 year.Example minute polymerizer 222 identifies family's sum of tuning minute (frame 804) from selected classification, and from selected classification identify exposure minute sum (frame 806).In other words, with the age be 45-54 year the male sex categorical match all families in, identify the sum of tuning minute of family and exposure minute.

The exemplary engine 224 that collects of Fig. 2 calculates for the probability collected (frame 808) based on above-mentioned summation.As mentioned above, can calculate according to the mode consistent with exemplary equation (4) probability collected by the exemplary engine 224 that collects.The exemplary engine 224 that collects calls randomizer to generate random or pseudo random number (frame 810), and if the random or pseudo random number obtained is less than or equal to probable value (frame 812), then the kinsfolk had in the family of media measuring instrument 108 is designated as spectators' (frame 814) of tuning fragment.On the other hand, at obtain random or pseudo random number is not less than or equals probable value, the kinsfolk had in the family of media measuring instrument 108 is not designated as spectators' (frame 816) of tuning fragment.

Return the frame 704 of Fig. 7, and continue hypothesis coupling all pay close attention to the family of classification number of thresholds be 30 (30), and this data pool fails to comprise the number of thresholds (frame 740) of qualified family, then exemplary independent distribution engine 218 is called with via independent allocation technique computes probable value (frame 710) by probabilistic engine 212.

Fig. 9 is exemplified with the example implementations of the independent distribution probability calculation (frame 708) of Fig. 7.In the example illustrated in Fig. 9, classification delimiter 226 identifies the whole group member families (such as, LPM, NPM etc.) (frame 902) with the set of the identical key forecast factor in donor pond.In addition, example categories delimiter 226 identifies the total tune minute of the respective numbers associated with the key forecast factor, and total family of the respective numbers associated with the key forecast factor exposes minute.As mentioned above, the key forecast factor can refer to the specific combination paying close attention to the concern age in sex and/or family in household size, family.Such as, classification delimiter 226 can identify the whole families with two or more kinsfolk (one of them is the male sex of age 45-54) in donor pond.In order to illustrate, assuming that example categories delimiter identifies 200 (200) the individual families with two or more member (one of them is the male sex of age 45-54).Also suppose that the number of combinations of the family (200) identified reflects 4500 tuning minute of total family and 3600 total exposures minute.

Except to except the viewing probability influential key forecast factor, one or more additional secondary predictor also can have impact to viewing probability.As mentioned above, market survey person can have the composite set or objective cross of paying close attention to classification, but having all those quantity paying close attention to the family of the composite set of classification is no more than threshold value (such as, 30 (30) individual families).But although can not represent the composite set paying close attention to classification fully from donor pond, the composite set of classification or the subdivision of objective cross can comprise the relatively large representative in donor pond.The independently subdivision (subgroup) that the composite set of classification is paid close attention in illustrative methods disclosed herein, equipment, system and/or goods identification and the corresponding family associated with each concern subgroup, it is applied to calculate family independently and is exposed probability.

Exemplary ratios manager 228 is from the identification of key forecast factor set and the quantity (such as, having 200 families of the scale of 2+ and the male sex of age 45-54) (frame 904) paying close attention to the family that subgroup is mated.From concern subgroup, exemplary ratios manager 228 identifies family's quantity of tuning minute and by this value divided by tuning minute of total family, with the tuning ratio of family (frame 906) calculated with pay close attention to subgroup and associate.Such as, if pay close attention to subgroup be tuned to same TV station (such as, WAAA) (such as, tuning TV station's classification) whole families, and 1800 tuning minutes are reflected in these families, then exemplary ratios manager 228 by 1800 divided by total family tuning minute 4500 to calculate the tuning ratio 0.40 (frame 906) of tv tuner platform classification family.Exemplary ratios manager 228 also identify exposure minute quantity and by this value divided by total exposure time, with the exposure ratio (frame 908) calculated with pay close attention to subgroup (such as, exemplary tv tuner platform classification) and associate.Such as, if pay close attention to subgroup be tuned to same TV station (such as, WAAA) (such as, family tuning TV station tolerance) whole families, and these families reflection 1320 exposure minute, then exemplary ratios manager 228 exposes ratio 0.37 (frame 908) divided by total exposure minutes 3600 to calculate tv tuner platform classification by 1320.If more can be paid close attention to subgroup (frame 910) from donor pond, then exemplary ratios manager 228 is selected next to pay close attention to subgroup (frame 912) and is controlled to turn back to frame 904.

After calculate the tuning ratio value of classification family and exposure ratio value for each concern subgroup, exemplary distribution engine 230 calculates the tuning ratio value of whole family and tuning minute of total family (such as according to concern classification, in this example 4500) long-pending (frame 914), and according to paying close attention to classification and calculate long-pending (frame 916) of all exposure ratio value and total exposure minute (such as, in this example 3600).Then final independent distribution probability can be calculated as exposing minute and family's ratio of tuning minute according to the mode consistent with exemplary equation (5).Such as, and as the description of above composition graphs 5, the exposure minute (17.47) of the expection obtained and the ratio of tuning minute of family (33.65) of expecting can be values 0.52.Ratio that this obtains instruction group member have the possibility of 52% be stay in three mouthfuls of families 45-54 year the male sex, be categorized into Old Age Homes, householder educates 9 years to graduating from the high school, and has two televisors in family, and on Monday watches TV station WAAA to Friday at 9:00AM to 12:00PM.

Visitor collects

Disclosed in above, personnel collect and whom utilize in the family and what family have viewed, and make for given tuning fragment, one or more kinsfolks can be designated and/or otherwise associate with exposure.But group member family has the visitor of the media be exposed in family, wherein effective visitor information is limited to age and sex.As mentioned above, exemplary PM comprises for the input end (such as, button) of each kinsfolk and for inputting for the age of any visitor mutual with media apparatus (such as, televisor) and the button of gender information.In view of collected group member family visitor information, illustrative methods disclosed herein, equipment, system and/or the application model that manufactures a product determine quantity for the visitor of the family not adopting PM and corresponding age/gender.

Visitor disclosed herein collects some similarity presenting and collect with personnel, and if desired in following discloses with reference in Fig. 1 to Fig. 9.Such as, above disclosed personnel collect to collect with the visitor of following discloses utilize tuning and exposure information to specify tuning fragment and calculation exposure to the ratio of tuning minute.But visitor collects viewing/tuning ratio (it is the ratio of total visitor exposure to the tuning exposure of total family rank) and reflects the average counter that visitor exposes instead of probability.Figure 10 further illustrates the mode processing visitor information compared with kinsfolk's exposure information.

In the illustrative example of Figure 10, the information for First Family 1002, second family 1004 and the 3rd family 1006 presented respectively by concrete concern TV station tuning 12, the 15 and 18 minutes time (being determined by each family with MM device and PM device).Although the illustrative example of Figure 10 only comprises three families, this example is only used to illustrate, and can consider the family of any amount.The member of First Family 1002 is exposed to seven minutes in 12 total tunes minute, and this causes the viewing probability of 7/12 (58.3%).In the second family 1004, the first member is exposed to complete 15 minutes, and the second member is exposed to five minutes in the tuning duration, obtains viewing probability (15+5)/(15+15) (66.7%).In the 3rd family 1006, the first member of this family is exposed to complete eight minutes, obtains watching probability 8/8 (100%).The overall viewing probability for this example home is determined according to the mode consistent with equation 6.

equation 6

In the example of illustrative equation 6, HH refers to family, and shown in exemplary equation 7, the example data from Figure 10 is applied to equation 6.

= \frac{(7) + (15 + 5) + (8)}{(12) + (15 + 15) + (8)} = \frac{35}{50} = . 70

Equation 7

In the example of illustrative equation 7, for exemplary 25-34 year the male sex the concern family of demographics group there is viewing probability 0.70.But, below according to the mode consistent with exemplary equation 8, average visitor is calculated to the analysis of the visitor in identical concern family and watch ratio.

equation 8

Shown in exemplary equation 9, the example data from Figure 10 is applied to equation 8.

= \frac{(12 + 10) + (15) + (5)}{(12) + (15) + (8)} = \frac{42}{35} = 1.20

Equation 9

In the illustrative example of equation 9, for report 25-34 year the male sex the concern family of visitor, for each viewing time presenting average 1.20 minutes for tuning minute.

Figure 11 is the schematic diagram that the exemplary guest of Fig. 1 collects the example implementations of engine 112.The exemplary guest of Fig. 1 collects engine 112 and constructs according to instruction of the present disclosure, and comprises average visitor's parameter (AVP) counter 1102, distribution engine 1104, randomizer 1106 and visitor and specify device 1108.As mentioned above, one or more the parts that the operation that exemplary guest collects engine 112 can collect engine 110 with Fig. 1 and Fig. 2 exemplary combine.In operation, exemplary people meter's interface 202 obtains the PM data associated with visitor, and wherein PM data are the PM devices from being arranged in the exemplary MMPM family 106 with both MM device and PM device.As described in conjunction with exemplary equation (1) above, exemplary PM interface 202 is identified in the visitor's data collected by threshold time section when the day before yesterday, to attempt according to its relative age these data weightings.

Exemplary guest collects engine 112 and calls example classes device 206 and/or example categories delimiter 226, classifies to the PM visitor's data obtained to pay close attention to classification based on one or more.As mentioned above, for given concern classification, identify the concrete family with these category associations.According to the family whether depositing number of thresholds in the donor pond of the visitor's data matched with the concern classification all expected, exemplary AVP counter 1102 will call the AVP computing technique of correspondence.Such as, if existence has the family exceeding number of thresholds of the concern classification of expectation (such as, 30 families), unit sorting technique then can be used to calculate AVP, independently sorting technique may be used for calculating AVP simultaneously, such as in conjunction with the independently category scheme that exemplary diagram 5 describes.

Exist in the event of the family of number of thresholds for paying close attention to the given set of classification, exemplary AVP counter 1102 is according to the mode consistent with exemplary equation 8 and calculate AVP as shown in figure 12.In the example of illustrated Figure 12, the classification paid close attention to comprises concrete tuning characteristic 1202 (such as, the family of Disney (Disney) platform is on Monday watched between 12:30 to 5:00PM to Friday) and concrete family's characteristic 1204 (such as, there is the family being in aged family life stage of 2 televisors).In addition, the visitor for two types performs the exemplary analysis of Figure 12: a kind of with 6-11 year women (row 1206) be associated, one and 55-64 year the male sex (row 1208) associate.In the family that the set of the expectation with presented concern characteristic (as determined by collected PM visitor's data) matches, exist 3892 minutes 6-11 year women visitor exposure (unit 1210) and 3109 tuning minute of total families (unit 1212).For this 6-11 year women visitor, the application of exemplary equation 8 produces AVP1.252 (unit 1214).In addition, with pay close attention to family that the expectation set of characteristic matches present 1081 minutes 55-64 year the male sex visitor expose (unit 1216), and always family's tuning minute (unit 1218) is retained in 3109 equally.For this 55-64 year the male sex visitor, the application of exemplary equation 8 produces AVP0.348 (unit 1220).

On the other hand, family number of thresholds for pay close attention to the disabled situation of expectation classification under (such as, be less than 30 families), after composition graphs 5 described above determines that desired exposure minute and tuning minute of expection are as classification ratio, exemplary AVP counter 1102 calculates AVP according to the mode consistent with equation 8.Figure 13 is exemplified with the exemplary tuning data of the target demographic for 6-11 year women and exposure data, the number of thresholds wherein meeting the family of the category combinations paid close attention to is disabled (such as, life stage=aged family+televisor number=2).In the example illustrated in Figure 13, reflection classification " life stage=aged family " family present 6-11 year women 443940 visitors exposure minute (unit 1302) and 733317 tuning minutes (unit 1304), and reflect the family of classification " televisor number=2 " present 6-11 year women 150844 visitors expose minute (unit 1306) and 285877 tuning minutes (unit 1308).In addition, 6-11 year women visitor exposure minute total amount present 1741474 minutes (unit 1310), and family's total amount of tuning minute presents 8200347 minutes (unit 1312).

Exemplary AVP counter 1102 and/or exemplary distribution engine 230 calculate for the exposure ratio 1314 of each concern classification and the tuning ratio 1316 for each concern classification.Continuing the illustrative example of Figure 13, is the ratio that visitor to expose minute to always to watch minute with the exposure ratio of life stage category associations, to produce scale factor 0.255 (result 1318).In addition, be 0.087 (result 1320) with the exposure ratio of TV televisor number category associations.The ratio being tuning minute of family to total tune minute with the exemplary tuning ratio of life stage category associations, to produce tuning ratio 0.089 (result 1322), and with the tuning ratio 0.035 (result 1324) of televisor number category associations.Although the example illustrated in Figure 13 comprises two (2) individual concern classifications, illustrative methods, equipment, system and/or manufacture a product and can comprise the concern classification of any amount.

Desired exposure minute value (unit 1326) is calculated as total exposure minute (unit 1310) and the quantity based on paid close attention to classification and the amassing of exposure ratio that any amount that occurs calculates (such as, as a result 1318 and result 1320) by exemplary AVP counter 1102.Exemplary AVP counter also by expect tuning minute value (unit 1328) be calculated as total tune time (unit 1312) and the quantity based on paid close attention to classification and tuning ratio that any amount that occurs calculates (such as, as a result 1322 and result 1324) long-pending.According to the mode consistent with exemplary equation 8, exemplary AVP counter 1102 calculates AVP value (unit 1330), and this AVP value is used to determine the quantity of visitor and relevant age, will describe in further detail below.

In order to determine the quantity of visitor and corresponding age, illustrative methods disclosed herein, equipment, system and/or goods have employed distributed model.Although the type of distributed mode described below is Poisson (Poisson) distribution, this distribution is object for example and not restriction.Poisson distribution is the discrete type probability distribution for representing their probability when known to the average ratio of a determined number event, and is employed the quantity (AVP before calculated is known average ratio) of the visitor of the tuning fragment of specifying viewing given in this article.The probability of Poisson distribution defines according to the mode consistent with exemplary equation 10.

p (v) = \frac{(λ_{d}^{v} * e^{- λ_{d}})}{v!}

Equation 10

In the example of the equation 10 illustrated, v represents the quantity of visitor, and p (v) represents the probability calculated for " v " individual visitor, and λ _drepresent given the AVP paying close attention to demographics group (such as, the women in age 6-11 year).Exemplary distribution engine 1104 defines the distribution of the exemplary Poisson distribution such as, and calculated candidate quantity pay close attention to the probable value of visitor, as being shown specifically further in Figure 14.

In the example illustrated of Figure 14, exemplary distribution engine 1104 for the first paid close attention to demographics group 1404 (such as, the women in age 6-11 year) have selected ten one (11) individual different visitor's quantitative values 1402, and exemplary distribution engine 1104 have selected ten one (11) individual different visitor's quantitative values (1406) for the second paid close attention to demographics group 1408 (such as, the male sex in age 55-64 year).For visitor's quantitative value of various discrete, exemplary distribution engine 1104 calculate corresponding probable value (see with age 6-11 year the row 1410 that is associated of women, and see with age 55-64 year the row 1412 that is associated of the male sex).Exemplary distribution engine 1104 also calculates the corresponding accumulated probability c (v) in each demographics group paid close attention to (see the row 1414 be associated with the women in age 6-11 year, and see the row 1416 be associated with the male sex in age 55-64 year).For convenience's sake, the exemplary cumulative distribution of Figure 14 allows to arrange probability between zero and the border of, makes exemplary randomizer 1106 to identify Query Value.

For the demographics group that each is paid close attention to, exemplary guest specifies device 1108 to call randomizer 1106 to generate random number, and when being quoted relative to cumulative distribution value, this random number discloses and belongs to the quantity that this institute pays close attention to the visitor of demographics group.Such as, if randomizer for age 6-11 year women's first group 1404 of being associated generate value 0.757000, then this value is specified device 1108 to associate to fall into value by exemplary guest is in the visitor (v) of 2.In addition, if randomizer for age 55-64 year the male sex's second group 1408 of being associated generate value 0.52700, then this value is specified device 1108 to associate to fall into value by exemplary guest is in the visitor (v) of 1.As a result, first group 1404 is considered to have two visitors, and each visitor has the age between 6-11 year, and second group 1408 is considered to have a visitor, and this visitor has the age between 55-64 year.Exemplary randomizer 1106 again by use to specify randomly age 6-11 year between two visitors of first group 1404 in the corresponding age of each, and specify randomly age 55-64 year between age of the visitor of second group 1408.Although above-mentioned example pays close attention to demographics group for being target with the women in age 6-11 year and the male sex in age 55-64 year and carrying out, but identical process can be repeated for all paid close attention to demographics groups, possibly other visitors are assigned to given tuning fragment.

The program 1500 of Figure 15 starts at frame 1502, and wherein exemplary PM interface 202 obtains and identifies and the data that the visitor of the visitor button that have selected the group member family in interest region (such as, DMA) is associated.Exemplary weights engine 210 according to based on from gather donor data date since time quantum pro rata to gather visitor's market demand weight (frame 1504).As mentioned above, more recent on time exponential quantity data point is usually more near exponential quantity 1.00 (see Fig. 3).In other words, when the image data of correspondence is more recent, the accuracy of rating index is better.

When analyzing paid close attention to market, example classes device 206 based on one or more pay close attention to classification obtained PM data classified.As mentioned above, pay close attention to classification and can include, but is not limited to the quantity (families etc. of three or more the platform televisors that such as, there is the family of a televisor, there is the family of 2-3 platform televisor, have) of the televisor paid close attention in paid close attention to age/gender combination, the particular home scale paid close attention to, the specific division of life span paid close attention to, the certain station/alliance/type paid close attention to, the specific time period paid close attention to, family and/or the education degree of householder.Although relatively a large amount of MMPM families 106 by have in above-mentioned classification at least one, in market survey process, for market survey, the MMPM family 106 of lesser amt will represent the objective cross of all classifications paid close attention in fact.

If exemplary guest collects engine 112 determine that the number of thresholds of the family be associated with the feature collection of preference and/or expectation is satisfied (such as, threshold value is at least 30 families) (frame 1508), then calculate AVP value (frame 1510) by exemplary AVP counter 1102 according to the mode consistent with Figure 12.On the other hand, when exemplary guest collect engine 112 determine that the number of thresholds of family is not satisfied (frame 1508), calculate AVP value (frame 1512) by exemplary AVP counter 1102 according to the mode consistent with Figure 13.Especially, exemplary AVP counter 1102 and/or exemplary distribution engine 230 pay close attention to classification calculation exposure ratio for each, and calculate tuning ratio for each classification of paying close attention to.The long-pending of each classification calculated distinctive exposure ratio and total exposure minute produces the exposure minute expected, and the long-pending of the distinctive tuning ratio of each classification calculated and total tune minute produces tuning minute that expects.The desired exposure that obtains minute and expection within tuning minute, be applied to exemplary equation 8 to produce corresponding AVP value.

Exemplary distribution engine 1104 defines distributed model to apply (such as Poisson distribution) (frame 1514), and according to any amount (v) calculating probability (frame 1516) of the mode consistent with exemplary equation 10 for paid close attention to visitor.Such as, Figure 14 exemplified with from zero (0) to ten (10) ten one (11) individual different visitor's quantitative values.Exemplary distribution engine 1104 also calculates accumulated probability, makes it possible to from the value be constrained between zero (0) and (1), select the selection from distribution.Exemplary distribution engine 1104 calls randomizer 1106 to select corresponding visitor's quantity (v) (frame 1520) in the accumulated probability set of the demographics set paid close attention to from each.Once each demographics set paid close attention to is provided with the quantity visitor determined, then select randomly to define age value be associated with tuning minute (frame 1522) for each visitor.

Environment is tuning

As mentioned above, can be physically connected to media apparatus (such as, televisor) and need the PM device of the installation of specialty to compare with employing, employing MM instead of PM characterizes the cost savings that home media exposure behavior is conducive to essence.Such as, MM can be mailed to group member, just can open without the need to specialty installation and/or the electrical equipment (such as, the media electrical equipment of such as DVD player, Set Top Box, televisor etc.) without the need to being connected to group member and work.Although use MM and do not use PM to save in fact the expense of the family of group member, some family has two or more media apparatus, these media apparatus are placed in relatively close room, wherein reach the room at the second media apparatus place from the sound of the first media apparatus, vice versa.In this case, the MM device in the first room may overflow based on the audio frequency be associated with the second media apparatus in the second room and record mistakenly and expose minute (vice versa).When MM device records exposure minute mistakenly, the tuning estimation of one or more family and/or supposition excessively can be reported or are expanded.Illustrative methods disclosed herein, equipment, system and/or goods are distinguished the tuning example of environment (such as, owing to overflowing) and truly tuning example.

Figure 16 is the schematic diagram of the exemplary realization of the tuning engine of exemplary environments 120 of Fig. 1.The tuning engine of exemplary environments 120 of Fig. 1 constructs according to instruction of the present disclosure.Illustrate in example at Figure 16, the tuning engine 120 of environment comprises PM interface 202 above disclosed in composition graphs 2 and MM interface 204.In addition, the example of the Figure 16 illustrated comprises simultaneous tuning watch-dog 1602, record management server 1604, platform comparer 1606, tuning type specify device 1608, Modeling engine 1614, code stack-manager 1616, automatic growth control (AGC) watch-dog 1610 and code to there is manager 1612.

In operation, exemplary PM interface 202 and exemplary MM interface 204 are from comprising available data pool (such as in interest region 104, LPM family, NPM family etc.) MMPM family 106 and MMH family 108 (the group member family such as, in direct marketing region (DMA)) collect family's tuning data.It is environment or real that the tuning engine of exemplary environments 120 calls exemplary simultaneous tuning watch-dog 1602 with the example identifying from the simultaneous tuning minute of the family data collected." simultaneous tuning " used herein refers to that in one family, two or more measuring instruments detect same media (such as, same TV station).In order to illustrate, assuming that detect WAAA TV station close to a MM of the first televisor in the first room, and also detect WAAA TV station close to the 2nd MM of the second televisor in the second room.May real a kind of possibility be these two media apparatus (such as, televisor) be unlocked and be tuned to WAAA TV station.But, another kind of possibility be the first televisor be unlocked and be tuned to WAAA TV station, and the second televisor is unlocked and when quiet be tuned to another TV station.Also have a kind of possibility be the first televisor be unlocked and be tuned to WAAA TV station, but the second televisor is not unlocked.In these cases, 2nd MM device may detect from the first televisor audio frequency (such as, overflow), and therefore make the media exposure (such as, consuming) be associated with the second televisor and/or kinsfolk measure to expand inadequately.

In some instances, record management server 1604 identifies the time quantum (such as, minute) of MM device record one TV station, and exemplary comparer 1606 determines whether the AP device matched with this MM device also have recorded identical TV station.If so, then exemplary tuning type specifies device 1608 to specify tuning minute to be accordingly real.On the other hand, if exemplary record manager 1604 identify MM device have recorded a TV station minute (such as, the code embedded is detected by MM device, the code embedded is transmitted by MM device and in aftertreatment, detected by the tuning engine 120 of environment during signature aftertreatment), and exemplary comparer 1606 determines that the AP device matched is not tuned to same TV station, then exemplary comparer 1606 determines whether the independent measurement mechanism in this family is tuned to same TV station (another AP and/or the MM device be associated with the second televisor in the second room in such as this family).If so, then this family within tuning minute, be considered to and/or be labeled as environment tuning/overflow (this should be left in the basket to prevent unsuitable excessive performance).On the other hand, when exemplary comparer 1606 determines do not have other measurement mechanism to be also tuned to same TV station in this family, exemplary tuning type specifies device 1608 to specify described minute for nonresonant.Exemplary simultaneous tuning watch-dog 1602 continues assessment from each tuning minute of receiving the data pool that exemplary group member family 104 collects.

In order to develop a kind of random device to determine the appearance of spilling (wherein, the model coefficient derived is derived for MMH family 108), extra predictive variable collected by the tuning engine of exemplary environments 120, and whether described predictive variable instruction spilling occurs.The model that described predictive variable is applied to such as regression model is conducive to calculating the coefficient/parameter whether occurring the probability overflowed in MMH family 108 to generate.Existence (the duration of the fragment of such as final publisher's audio code (FDAC) and collection that at least three predictive variables whether occurred comprise automatic growth control (AGC) value, embedded code is overflowed in instruction.

Generally speaking, by comparing the AGC value (such as, calculating the difference between AGC value) between the different MM device of in a family two, the instruction of spilling can be assessed.The MM device relatively arranged near the first televisor such as has low AGC value more possibly, this is because, compared with the AGC value be associated with the sound from televisor relatively far away, there is relatively higher volume.AGC value is set up by acoustic gain circuit usually, acoustic gain circuit attempt to identify and/or detect there is the acoustic energy of relatively low volume time than applying larger gain (such as, amplifying) when attempting to detect the acoustic energy with higher volume.Volume may such as due to send the larger distance in source of acoustic energy and lower.In addition, the quantity of the code that each chronomere detects and/or density are the extra example predictive variablees that can be applied to model, to derive the instruction of the appearance whether possibility of spilling.The duration of fragment is another predictive variable useful in the instruction of overflowing, and as will be described in further detail below, but is not limited thereto.

What each collected by the exemplary AGC watch-dog 1610 of Figure 16 minute is assigned to corresponding AGC value.What the exemplary codes of Figure 16 existed that manager 1612 collects to each minute specifies the designator whether corresponding with the existence of embedded code.In some instances, code detection activity can occur in the last handling process of the original audio information of measuring instrument collection.In other example, code is by real time or close to detecting in real time.The tuning engine 120 of exemplary environments of Figure 16 is based on whether embedded code being detected and being separated by the example of simultaneous tuning minute.Such as, the dependent variable that occurs with the true or ambient condition determined before representing of Modeling engine 1614 is to prepare regression model.Exemplary AGC watch-dog 1610 is for minimum (such as, minimum) AGC of the audio data sets determination home devices of particular monitored period and/or collection.For each device and minute, what exemplary AGC watch-dog 1610 gathered about minimum AGC value and each minute determines AGC difference.

There is manager 1612 and identify for the one in the type of the code in the collection MM data of the device (such as, televisor, radio etc.) in family and three kinds of possibility scenes of existence in the exemplary codes of Figure 16.The first may scene be all nonexistent code in collection MM data for any one device of paid close attention to family.The second may scene be collect MM data to have some codes for some devices in family, but not all device all have collect minute in the correlative code that detects.The third may scene be for family collection MM data collect all data in all there is code.In other words, in each all device minute in this family collected of tuning data, there is corresponding code.

If any one measuring instrument in family collect minute in not there is the code of collection, then the exemplary simultaneous tuning watch-dog 1602 of Figure 16 is placed on larger weight on the segment durations for a type of this family.Such as, if televisor is tuned to WAAA platform, then the MM device near this televisor will have than being positioned at further from the MM device of the position of this televisor relatively longer Hoarding segment duration.From the intensity Possible waves being positioned at the sound sent further from the televisor of the position of same MM device, make this MM device may not capture complete segment durations.The exemplary simultaneous tuning watch-dog 1602 of Figure 16 is the longest (such as the identification of each family, maximum) segment durations, and calculate the duration that will be applied to the logistic regression matching of collecting data according to the mode consistent with exemplary equation (11) poor.

equation (11)

In the illustrated example of equation (11), model has the environment value or actual value that response (dependence) dependent variable is assigned to as each simultaneous tuning minute.Independent variable X ₁... X _kmodel coefficient B can be utilized ₁... B _kcoding and/or classification.Classification can be represented by random scale, the AGC value of the scope such as wherein with subgroup from zero to one hundred.

If some measuring instruments in family collect minute in have collected code (such as, the original audio collected, which is embedded code and be identified during voice data aftertreatment subsequently), but other measuring instrument is not collected, then the exemplary Modeling engine 1614 of Figure 16 uses with the data that AGC difference is associated according to the mode Modling model consistent with exemplary equation (11).If all measuring instruments in family collect minute in all have collected code, then exemplary codes stack-manager 1616 is determined to count for the maximum non-storehouse counting of home devices and maximum storehouse.As used herein, the Stacked codes example that refers to code reparation when code a part of being detected and/or collect.When MM device does not correctly collect whole code content, storehouse process fills up the code section be not detected.Generally speaking, relatively near the measuring instrument device (such as, MM) of media apparatus (such as, televisor) by due to the such as relative ability that there is closer to measuring instrument device better collection and do not need the non-Stacked codes repaired or fill.But when measuring instrument device is when the position work relatively far away apart from monitored device, the ability that described measuring instrument device collects whole code exactly becomes more difficult and mistake.Code stack-manager 1616 in illustrated example determines to be applied to the difference between storehouse count value between the measuring instrument device in the family of model and non-storehouse count value.In addition, the simultaneous tuning watch-dog 1602 in illustrated example identifies the maximum average number of seconds of the collection code for all measuring instrument devices in family, and calculates the difference between those home devices.Collect the difference of the number of seconds of code, Stacked codes counting is applied to the exemplary model of equation (11) to derive corresponding model coefficient (such as, B with non-Stacked codes count difference value and AGC difference ₁... B _k).

As mentioned above, there is manager 1612 and identify for the one in the type of the code in family and three kinds of possibility scenes of existence in the exemplary codes of Figure 16, and the various combination of applied forcasting variable (such as, the counting of AGC value, segment durations, the relatively non-Stacked codes of Stacked codes counting) is carried out based on the scene detected.Corresponding predictive variable is applied to the exemplary model of equation (11) by each in these scenes, and the exemplary Modeling engine 1614 of Figure 16 is according to the probability of the mode calculation overflow consistent with equation (12).

equation (12)

Based on obtained probable value and the threshold value such as set up by market researcher, it is tuning or truly tuning that each simultaneous tuning minute can be identified as environment.Such as, if probable value is more than or equal to 0.50, then within described minute, environment can be designated as tuning.On the other hand, for the probable value being less than 0.50, within described minute, can be designated as truly tuning.

The program 1700 of Figure 17 starts at frame 1702, and the exemplary PM interface 202 in wherein illustrated example and exemplary MM interface 204 collect tuning data from the MMPM family 106 in group member family 104 and MMH family 108.The exemplary simultaneous tuning watch-dog 1602 of Figure 16 identifies that the simultaneous tuning minute in these families is environment or real (frame 1704), describes in more detail below in conjunction with Figure 18.

Figure 18 comprises from the additional detail in the example illustrated in Figure 17.Identify simultaneous tuning minute be environment or real time, the exemplary record manager 1604 of Figure 16 identifies minute (frame 1802) of MM device (the MM device such as, in the MMPM family 106) scoring table in the family that pays close attention to.The platform comparer 1606 of Figure 16 determines whether the AP device in described paid close attention to family also records the platform (frame 1804) identical with this MM device at synchronization.In some instances, record management server 1604 compares with the timestamp minute to be associated collected from MM device with the timestamp minute to be associated that the PM device from the same family is collected.If described timestamp mates and the platform detected is identical, then the exemplary tuning type of Figure 16 specifies device 1608 to be appointed as corresponding minute truly tuning (frame 1806).Exemplary simultaneous tuning watch-dog 1602 determines whether there is extra minute (frame 1808) from paid close attention to family that will analyze.If so, then exemplary simultaneous tuning watch-dog 1602 is selected next minute to carry out analyzing (frame 1810) and is controlled to return frame 1804.

If exemplary the comparer 1606 of Figure 16 determine AP do not recording the platform identical with the MM device in family (may due to the multiple media presentation device in family (such as, televisor) be tuned to different platforms or be closed) (frame 1804), then exemplary comparer 1606 in illustrated example determines whether other device is tuned to identical platform (frame 1812).As mentioned above, illustrative methods disclosed herein, equipment, system and/or goods adopt the MMPM family 106 not only having had MM device but also had PM device, are eliminated to make the ambiguity of actual device behavior.Once create model coefficient based on this behavior observed in MMPM family 106, the data of collecting from MMH family 108 just can utilize described coefficient to estimate, to allow the instruction of calculation overflow.Like this, the group member family without PM can be effectively utilised.As a result, the group member family of larger quantity can be realized in exemplary interest region 104, and there is no the cost of the increase of PM device, relative to MM device, the installation of PM matching requirements specialty, relatively more substantial training and/or more daily servicing.

If exemplary comparer 1606 determines that other device in family is tuned to identical platform (frame 1812) (such as, detection based on same code), then exemplary tuning type specifies device 1608 corresponding minute to be appointed as environment tuning (herein also referred to as spilling) (frame 1814).On the other hand, if exemplary comparer 1606 determines that other device in family is not tuned to identical platform (frame 1812), then the exemplary tuning type in illustrated example specifies device 1608 corresponding minute to be appointed as nonresonant minute (frame 1816).

Return Figure 17, each minute is assigned to corresponding AGC value (frame 1706) by the exemplary AGC watch-dog 1610 in illustrated example.As mentioned above, the AGC value be associated with collected minute in some example predictive variablees contributes to the probability of the tuning appearance of computing environment.In addition, another example predictive variable discussed above comprise collected media minute in embedded code existence whether.It is specify the designator (frame 1708) whether relevant with the existence of embedded code in each minute that exemplary codes in illustrated example exists manager 1612.The example of simultaneous tuning minute is separated (frame 1710) based on such embedded code whether being detected by the tuning engine of exemplary environments 120, describes in more detail below in conjunction with Figure 19.

In the example illustrated in Figure 19 (frame 1710), the Modeling engine 1614 in illustrated example is to reflect that the dependent variable of truly corresponding or ambient condition designator is to prepare regression model (frame 1902).For paid close attention to each home devices, the minimum AGC value (frame 1904) of two or more MM devices determined by exemplary AGC watch-dog 1610, and determines the difference (frame 1906) between them.In view of the MM device in paid close attention to family may collect less than code, the code collecting some minutes and collect less than other minute code or collect the probability of code of all minutes, there is manager 1612 and identify that situation is suitable for (frame 1908) in the exemplary codes in illustrated example.

If there is manager 1612 in the exemplary codes in illustrated example identifies first category code wherein not detected, then maximum segment duration (frame 1910) of being associated with MM device determined by exemplary simultaneous tuning watch-dog 1602, and calculates the difference (frame 1912) between them.The exemplary Modeling engine 1614 of Figure 16 according to the mode consistent with equation (11) to the market demand logistic regression matching (frame 1914) of collecting, as mentioned above.Particularly, when family collect minute in any code do not detected time, the exemplary model of equation (11) is modified to consider (1) AGC value in collected segment durations and (2) difference (frame 1914).

If there is manager 1612 and identify and wherein some codes detected in some minutes and the second classification (frame 1908) code not detected in other minute in the exemplary codes in illustrated example, then the exemplary Modeling engine 1614 of Figure 16 according to the mode consistent with equation (11) to collected market demand logistic regression matching (frame 1916).But, in this application of exemplary equation (11), model adopt corresponding collect minute in (1) AGC value and (2) code existence whether (frame 1916).

If exemplary codes exist manager 1612 identify wherein all collections minute in the 3rd classification (frame 1908) of all codes detected, then the exemplary codes stack-manager 1616 of this example determines the code self whether complete (frame 1918) that is checked through.As mentioned above, although exemplary MM device can detect from media apparatus (such as, televisor) and/or catch the code that may be embedded in media, the quality of the code detected may be different.Such as, this species diversity may be because MM device collects audio frequency from the relative televisor away from this MM device position.Under these circumstances, one or more stack manipulation can utilize code data accurately to supplement the lost part of the code detected.The exemplary codes stack-manager 1616 of this example identifies the difference (frame 1920) of the quantity of the code of the relatively non-storehouse of the code about the storehouse detected between the MM device in family.In addition, exemplary simultaneous tuning watch-dog 1602 calculates the corresponding difference (frame 1922) between the average number of seconds (such as, maximum average value) of the code of each measuring apparatus in family and each measuring apparatus.Exemplary Modeling engine 1614 in illustrated example according to the mode consistent with exemplary equation (11) to collected market demand logistic regression matching (frame 1924).But, in this application of exemplary equation (11), the difference (frame 1924) between the average number of seconds of the difference between the embedded code of (1) AGC value that model have employed, (2) storehouse/non-storehouse and the code between (3) measuring apparatus.

Return Figure 17, the exemplary Modeling engine 1614 of this example will from model (such as according to the mode consistent with exemplary equation (12), equation (11)) coefficient that calculates is applied to probability calculation, to determine the tuning probability (frame 1712) that should be classified as spilling (environment is tuning) of given minute.

ON/OFF detects

As mentioned above, adopt MM can perform by random with characterization home media viewing behavior, instead of adopt PM to save the expense will paid for relatively costly PM device.When adopting MM device to collect sound signal (tuning) data from family, collect minute in some can comprise code (such as, collect in original audio and be passed to the embedded code of ON/OFF detecting and alarm 130 for aftertreatment), collect minute in some can via signature analysis (such as, collected by MM device and be passed to the analysis of original audio of ON/OFF detecting and alarm 130 for comparing with the audio signature of one or more signature database) and analyzed, and collect minute in some neither can have the code matched to media identification and also not there is corresponding signature.

Figure 20 exemplifies the exemplary record statistical form 2000 of the square frame of 24 (24) hours with the tuning data collected from the example MM device exemplary home.In the illustrative example of Figure 20, from family collect minute some parts relevant to code 2002, this also indicates the media apparatus in family (such as, televisor) to be opened.In addition, when compared with reference database, from family collect minute some parts relevant to the signature 2004 of the media detected, the signature 2004 of these media allows media identification.But, from family collect minute other parts also neither there is the code mated with the known media reference database 2006 not there is the signature mated with the known media in reference database 2006 yet.

Here by neither have can with the code that reference database uses do not have yet corresponding sign minute to be called all other tuning (AOT) minute.When having PM device, will in open mode (such as, power supply status ON based on the power state detection device of PM device) under detect media apparatus (such as, televisor), but do not have platform and/or media to be registered as to have tuning.In other cases, media apparatus can be in mute state or closed condition (such as, power supply status OFF), therefore can not be enough in the audio frequency of record by emission energy.Exemplary method disclosed herein, system, equipment and/or product application determine within AOT minute, to be relevant to closed condition or the random fashion relevant with open mode, this closed condition or open mode can be correlated with other media apparatus purposes (such as, video-game uses, uses meeting etc.) independent of media program.

Figure 21 training centre according to the present invention builds, the schematic diagram of the exemplary ON/OFF detecting and alarm 130 of Fig. 1.In the illustrative example of Figure 21, ON/OFF detecting and alarm 130 comprises PM interface 202, MM interface 204, AGC watch-dog 1610 and Modeling engine 1614 disclosed in composition graphs 12 above and Figure 16.

In operation, example PM interface 202 to be collected minute from the PM device (such as, active/passive people meter) in the family relevant to three classifications that media apparatus uses.Relevant with the certain station of the first category that minute relevant media apparatus that some are collected uses and the media that the mode such as by code or signatures match identifies or media.The second classification used to minute relevant media apparatus that other is collected and non-program be correlated with use situation (such as, video-game play, video conference activity, home photos viewing etc.) relevant.Relevant with the situation of the 3rd classification that minute relevant media apparatus that other is collected uses and media apparatus power-off.

Exemplary MM interface 204 is also from the MM device collection minute in family.As mentioned above, because MM device 204 is not connected to media apparatus physically, so MM interface 204 directly can not verify whether media apparatus starts shooting, but only collect the information based on audio frequency via one or more built-in microphone.Exemplary MM interface 204 can collect the minute data of scoring table or media, or minute is appointed as AOU by what collect.Exemplary AGC watch-dog 1610 collects AGC data from for each the exemplary MM interface 204 of each corresponding minute and exemplary PM interface 202, and exemplary Modeling engine 1614 prepares regression model with the data collected by matching in the mode consistent with example equation (13).

equation (13)

In equation (13) illustrative example, HUT instruction " family uses televisor " is opened (such as, ON power supply status), OFF indicates OFF power supply status, and independent variable (X) comprises AGC value, period information and/or from the number of minutes that there occurs code reader record.

Exemplary Modeling engine 1614 uses the coefficient (B) derived to calculate each minute for the probability opening (HUT) or close by the mode consistent with equation (14).

equation (14)

The program 2200 of Figure 22 starts at frame 2202, and at frame 2202, exemplary PM interface 202 to be collected minute from PM device, this PM device and scoring table minute, televisor using but do not have record minute, televisor power-off minute relevant.Exemplary MM interface 204 from the MM device dual platen family collect be recorded with platform minute with minute relevant minute (frame 2204) of AOU.The AGC value (frame 2206) relevant to each minute collected by exemplary PM interface 202 and MM interface 204 collected by exemplary AGC watch-dog 1610.

Exemplary Modeling engine 1614 carrys out preparation model (frame 2208) in the mode consistent with example equation (13) based on AGC value, period and minute quantity from last MM device record.This model can include but not limited to regression model, derives coefficient after the data wherein can collected in matching.The probability that exemplary the Modeling engine 1614 and open mode of home media device or relevant any specific of closed condition that uses the model coefficient derived to calculate is paid close attention to minute.These coefficients derived can relevant to the member family in the interest region 104 only with MM device 108 (frame 2210).

Figure 23 be can perform Fig. 6-9, Figure 15, Figure 17-19 and Figure 22 instruction with the block diagram of the example processor platform 2300 of the tuning engine 120 of the environment realizing Fig. 1, Fig. 2, Figure 11, Figure 16 and Figure 21, interpolation engine 110, visitor's interpolation engine 112 and ON/OFF detecting and alarm 130.Processor platform 2300 can be the calculation element of such as server, personal computer, internet device, digital VTR, personal video recorder, Set Top Box or other type any.

The processor platform 2300 of illustrative example comprises processor 2312.The processor 2312 of illustrative example is hardware.Such as, processor 2312 can realize by from one or more integrated circuit of the family of any expectation or manufacturer, logical circuit, microprocessor or controller.

The processor 2312 of illustrative example comprises local storage 2313 (such as, Cache).The processor 2312 of illustrative example via bus 2318 with comprise the primary memory of volatile memory 2314 with nonvolatile memory 2316 and communicate.Volatile memory 2314 can be realized by the random access storage device of Synchronous Dynamic Random Access Memory (SDRAM), dynamic RAM (DRAM), RAMBUS dynamic RAM (RDRAM) and/or other type any.Nonvolatile memory 2316 can be realized by the memory storage of flash memory and/or other desired type.The access to primary memory 2314,2316 is controlled by memory controller.

The processor platform 2300 of illustrative example also comprises interface circuit 2320.Interface circuit 2320 can be realized by the interface standard of such as any type of Ethernet interface, USB (universal serial bus) (USB) and/or PCI express interface.

In illustrative example, one or more input media 2322 is connected to interface circuit 2320.One or more input media 2322 allows user data and order to be input in processor 2312.Such as, one or more input media can be realized by audio sensor, microphone, camera (static or video), keyboard, button, mouse, touch-screen, tracking plate, tracking ball, mousegrid (isopoint) and/or speech recognition system.

One or more output unit 2324 is also connected to the interface circuit 2320 of illustrative example.Such as, output unit 2324 can be realized by display device (such as, light emitting diode (LED), Organic Light Emitting Diode (OLED), liquid crystal display, cathode-ray tube display (CRT), touch-screen, light emitting diode (LED), printer and/or loudspeaker).Therefore, the interface circuit 2320 of illustrative example generally includes graphics driver card, graphdriver chip and graphdriver processor.

The interface circuit 2320 of illustrative example also comprises the communicator of such as transmitter, receiver, transceiver, modulator-demodular unit and/or network interface unit to facilitate via network 2326 (such as, Ethernet connection, Digital Subscriber Line (DSL), telephone wire, concentric cable, cell phone system etc.) exchange data with external mechanical (such as, the calculation element of any kind).

The processor platform 2300 of illustrative example also comprises one or more mass storage device 2328 for storing software and/or data.The example of such mass storage device 2328 comprises floppy disk, hard disk drive, CD drive, blu-ray disc drives, RAID system and digital universal disc (DVD) driver.

The instruction 2332 of the coding of Fig. 6-9, Figure 15, Figure 17-19 and Figure 22 can be stored in the detachable tangible computer readable memory of mass storage device 2328, volatile memory 2314, nonvolatile memory 2316 and/or such as CD or DVD.

From above, what it will be appreciated that is, disclosed method, equipment and goods allow audience measurement technology to occur together with more substantial in fact family above, wherein, each family is by adopting based on the code reading device of audio frequency instead of relatively more expensive people meter's device to have measurement mechanism cost lower in fact.Example disclosed herein allows the behavior probability determining can be applicable to family, and this family does not have people meter's device but adopts the media measuring instrument gathering audio frequency when not needing to be physically connected to media apparatus (such as, televisor).Such example allows to make for carrying out behavior probability calculating based on other family comprising people meter's device, and wherein, this calculating discloses behavior probability with the random fashion of the expectation value of Corpus--based Method importance.

Exemplary method disclosed herein, system, equipment and/or goods are also convenient to the random fashion determining the probability not adopting the environment in the family of people meter's device tuning.Here in discloseder examples, adopt group member's audience measurement instrument (such as, people meter) and media measuring instrument (such as, gathering audio frequency when not needing to be physically connected to media apparatus) the two to obtain data relevant with one or more automatic growth control (AGC) value to media codes state.Based on the code status obtained and AGC value, example disclosed herein is to determine the mode model of creation coefficient supporting the probability that the environment of the expectation value of statistical significance is tuning, and this model coefficient can be applied to the family only with media measuring instrument.In addition, disclosed herein is with AGC value about and the data that obtain to be used from computation model coefficient with period information one, this model coefficient instruction media apparatus (such as, televisor) is energising or power-off.

The quantity of visitor and the probability at corresponding age thereof in other exemplary method disclosed herein, system, equipment and/or product marking family.Particularly, example disclosed herein is based on exposure minute and within tuning minute, calculate average visitor's parameter (AVP), this exposure minute and be applied to Poisson distribution for tuning minute further to determine the probability of the visitor in family with specific quantity.Such probability consider have age-specific reference range pay close attention to target demographic, this age-specific reference range can be selected based on the input from randomizer.

Although there has been disclosed particular example method, equipment and goods, the coverage of this patent is not limited thereto.On the contrary, this patent covers all methods, equipment and the goods that quite fall in the scope of claims of this patent.

Claims

1. a method for computing medium installation's power source shape probability of state, described method comprises:

2. method according to claim 1, wherein said power supply status probability calculation comprises described model coefficient, period information and the AGC value from described 2nd MM in described second family.

3. method according to claim 1, described method also comprises with the independent variable computation model coefficient based on minute quantity from the MM scoring table in described First Family.

4. method according to claim 1, described method also comprises multiple minutes that identification does not comprise code or the signature mated with reference database.

5. method according to claim 4, wherein said power supply status probability calculation determines it within described multiple minutes, is associate with pass power supply status or associate with other tuning states all.

6. an equipment for computing medium installation's power source shape probability of state, described equipment comprises:

7. equipment according to claim 6, wherein said Modeling engine is used for described power supply status probability calculation by comprising described model coefficient, period information and the AGC value from described 2nd MM in described second family.

8. equipment according to claim 6, wherein said Modeling engine is by with the independent variable computation model coefficient based on minute quantity from the MM scoring table in described First Family.

9. equipment according to claim 6, described equipment also comprises detecting and alarm, and it identifies multiple minutes that do not comprise code or the signature mated with reference database.

10. equipment according to claim 9, wherein said Modeling engine will based on described power supply status probability calculation determine within described multiple minutes, be with pass power supply status associate or with other tuning (AOT) state relations all.

11. 1 kinds of tangible machine readable storage medium storing program for executing comprising instruction, described instruction causes machine at least upon being performed:

12. storage mediums according to claim 11, wherein said instruction also causes described machine to comprise described model coefficient, period information and the AGC value from described 2nd MM in described second family when calculating power supply status probability upon being performed.

13. storage mediums according to claim 11, wherein said instruction also causes the independent variable computation model coefficient of described machine based on minute quantity from the MM scoring table in described First Family upon being performed.

14. storage mediums according to claim 11, wherein said instruction also causes described machine recognition not comprise multiple minutes of code or the signature mated with reference database upon being performed.

15. storage mediums according to claim 14, wherein said instruction also causes described machine to determine it within described multiple minutes, is associate with pass power supply status or associate with other tuning states all based on described power supply status probability calculation upon being performed.