JP2011150380A - Viewer information estimation device and method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a viewer information estimation device for estimating a viewer in an arbitrary time. <P>SOLUTION: A viewer estimation device 10 is provided with a time zone-specific viewing feature storage part 1 for storing time zone-specific viewing feature data which are analyzed on the basis of operation history of a device for viewing, and which show at least one feature related to a program viewed in each time zone and a numerical value corresponding to the feature; a family feature storage part 2 for storing family feature data quantitatively showing the features of the viewing content of each of the constituents of the family; and a similarity calculation part 3 for extracting the time zone-specific viewing feature data in the designated time zone, and for calculating the similarity of the time zone-specific viewing feature data and the family feature data. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an apparatus for estimating a family who has watched a program on a home television.

  With the spread of digital broadcasting and IP (Internet Protocol) broadcasting, it has become possible to provide useful information to viewers. In addition, a recommendation technique for distributing information suitable for a user's preference is used in shopping sites on the Internet. However, since television is often shared by multiple families, it is necessary to identify the viewer of the TV at that time in order to send appropriate information to the viewer. The technology has been devised.

  Japanese Patent Application Laid-Open No. 2004-133620 gives a program that recommends viewers based on history information related to the identifiers by assigning different identifiers to a plurality of input devices, specifying the viewers based on the identifiers of the input devices used for the input operation. An apparatus for determining is disclosed.

  Japanese Patent Laid-Open No. 2004-228688 discloses content that identifies a viewer by performing user authentication using a remote controller, an RF-ID tag, or the like, or estimates a viewer from a viewing history, and corresponds to the viewer based on the viewing history. An apparatus for supplying is disclosed.

JP 2007-251848 A JP 2007-189397 A

  In the device of Patent Document 1, the viewer is specified based on the identifier of the input device used for the input operation. For this reason, there is a problem in that the user has to use different input devices for each user, which is inconvenient.

  Also in the apparatus of Patent Document 2, when user authentication is performed using an RF-ID tag or a remote controller, the user needs to hold the tag over the detection unit or perform a predetermined remote control operation, which is not convenient. There was a problem. There is also a problem that a device for recognizing a user, such as an RF-ID tag or a camera, must be added to the system.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide a viewer information estimation device that can estimate a viewer at an arbitrary time without reducing convenience.

  Another object of the present invention is to provide a viewer information estimation device capable of estimating a viewer at an arbitrary time with a simple system configuration.

  The present invention is based on the operation history of the viewing apparatus and stores time-specific viewing feature data representing at least one feature related to a program viewed in each time zone and a numerical value corresponding to the feature. A viewing feature storage unit for each band, a family feature storage unit for storing family feature data that quantitatively represents the feature of viewing content for each member of the family, and a viewing feature data for each specified time zone. And a similarity calculation unit that calculates the similarity with the family feature data.

  The present invention relates to at least one feature relating to a program watched in each time zone, which is analyzed based on the operation history of the viewing device, and viewing feature data by time zone representing a numerical value corresponding to the feature, and family members Family feature data that quantitatively represents the features of each viewing content is created, viewing feature data for each specified time zone is extracted, and similarity with the family feature data is calculated This is a viewer estimation method.

  The present invention relates to at least one feature relating to a program that is analyzed based on an operation history of a viewing device and watched in each time zone, and viewing feature data by time zone that represents a numerical value corresponding to the feature, and members of a family A similarity calculation function that extracts viewing feature data by time zone in a specified time zone and calculates similarity with the family feature data in a computer using family feature data that quantitatively represents the feature of each viewing content It is a program for realizing.

  According to the present invention, viewers at an arbitrary time can be estimated without reducing convenience.

FIG. 1 is a schematic diagram of an embodiment of the present invention. FIG. 2 is a block diagram showing a functional configuration of the viewer estimation apparatus according to the embodiment of the present invention. FIG. 3 is a diagram showing an example of the operation history database. FIG. 4 is a diagram illustrating an example of the operation ID. FIG. 5 is a diagram showing an example of a program feature database. FIG. 6 is a diagram showing an example of an electronic program guide. FIG. 7 is a diagram showing an example of viewing characteristic data by time zone. FIG. 8 is a diagram showing an example of a family feature database. FIG. 9 is a flowchart for explaining the processing operation in the scene A in which the operation history is accumulated in the operation history database. FIG. 10 is a flowchart for explaining the processing operation in the scene B in which the viewing feature data by time zone is registered in the viewing feature storage unit by time zone. FIG. 11 is a flowchart for explaining the processing operation in the scene A in which the operation history is accumulated in the operation history database. FIG. 12 is a flowchart for explaining the processing operation in the scene C in which the viewer estimation degree is calculated. FIG. 13 is a diagram for specifically explaining the calculation of the similarity. FIG. 14 is a flowchart for explaining the processing operation in the scene C in which the viewer estimation degree is calculated. FIG. 15 is a diagram showing an example of an operation history database in a modification of the present embodiment. FIG. 16 is a flowchart for explaining the processing operation of the viewer estimation apparatus according to the modification of the present embodiment.

  Embodiments of the present invention will be described below with reference to the drawings.

  First, an outline of an embodiment of the present invention will be described with reference to FIG.

  A viewer estimation device 10 according to an embodiment of the present invention includes a viewing feature storage unit 1 for each time zone, a family feature storage unit 2, and a similarity calculation unit 3.

  The time-based viewing feature storage unit 1 is analyzed based on an operation history performed to view a program, and represents at least one feature related to the program viewed in each time zone and a numerical value corresponding to the feature. The band-specific viewing feature data is stored.

  The family feature storage unit 2 stores family feature data that quantitatively indicates the feature of the viewing content for each member of the family.

  The similarity calculation unit 3 extracts the viewing feature data 1 for each designated time zone and calculates the similarity with the family feature data 2 for each family member. This calculated value is a viewing estimation degree for each family member at the designated time, and is a value indicating the degree of being considered a viewer at the designated time.

  Details of embodiments of the present invention will be described with reference to the drawings. FIG. 2 is a block diagram showing the configuration of the viewer estimation system according to this embodiment. This viewer estimation system includes a viewer estimation device 10, an operation device 11, and an estimated viewer information acquisition device 12.

  The viewer estimation device 10 includes an operation history generation unit 111, an operation history database 112, an operation history determination unit 113, a program feature database 114, a viewing feature analysis unit 115, a viewing feature storage unit by time zone 116, A viewer information estimation unit 117, a similarity calculation unit 118, and a family feature database 119 are provided.

  The time zone viewing feature storage unit 116 is the time zone viewing feature storage unit 1 of FIG. 1, the family feature database 119 is the family feature storage unit 2 of FIG. 1, and the similarity calculation unit 118 is the similarity of FIG. Each corresponds to the degree calculation unit 3.

  The operation device 11 is a device for a viewer to perform various operations for viewing a program, and includes a television remote controller. The operation device 11 transmits information on the operation content input by the viewer.

  The estimated viewer information acquisition device 12 is a device for designating and inputting a day of the week and a time when the viewer wants to know. The estimated viewer information acquisition device 12 transmits the input data of the designated day of the week and time to the viewer information estimation unit 117. Further, the estimated viewer information acquisition device 12 receives and outputs the viewing estimation degree from the viewer information estimation unit 117.

  The operation history generation unit 111 receives the operation content information transmitted from the operation device 11, generates operation history data based on the operation content, and accumulates and stores the operation history data in the operation history database 112.

  An example of the operation history database 112 is shown in FIG. The operation history data stored in the operation history database 112 includes, for example, an operation history ID, an operation date and time, a content ID of a viewed program, an operation ID indicating the type of operation, and whether or not the operation data is unprocessed. It includes data such as an unprocessed flag. An example of the operation ID is shown in FIG.

  Note that the operation history generated by the operation history generation unit 111 is not limited to the operation of a single television, and may include operations on various devices (viewing devices) for viewing a program. In addition to the television, the viewing device includes a set-top box connected to the television, a hard disk recorder, a recording / playback device such as a DVD / Blu-ray recorder, and a game machine. The operation device 11 includes an input operation device corresponding to each device.

  The operation history determination unit 113 reads the operation history data stored in the operation history database 112, determines whether or not the operation history data is valid for estimating the viewer, and views the operation history data determined to be valid. This is sent to the feature analysis unit 115. In the present embodiment, it is assumed that operation history data effective for estimating a viewer includes information on content to be viewed in the operation history. For example, since the operation history of turning on / off the television does not include content information, it is determined that the operation history is not valid. In addition, the operation history such as fast forward and rewind may be determined to be valid because the content to be viewed can be identified.

  The program feature database 114 stores program feature data that quantitatively represents the features of each program. An example of the program feature database 114 is shown in FIG. The program feature data includes data such as a content ID, a keyword, and a weight. Of these keywords, keywords that are characteristic from the electronic program guide illustrated in FIG. 6 are extracted and registered. As a method for extracting a keyword to be a feature, a technique that can extract a word to be a feature in a sentence such as a tf-idf method is used. As the weight, a value that quantitatively represents the degree of feature of the keyword in the program is set. For example, when the tf-idf method is used for the process of extracting the keyword from the electronic program guide, the calculated tf-idf value may be set as the weight.

  The viewing feature analysis unit 115 performs analysis using the operation history data received from the operation history determination unit 113, and reflects the analysis result in the viewing feature data classified by time zone stored in the viewing feature storage unit sorted by time zone. Specifically, the viewing feature analysis unit 115 reads program feature data corresponding to the viewing content indicated by the operation history from the program feature database 114. Also, the viewing characteristic data for each time zone corresponding to the operation date and time indicated by the operation history is read from the viewing characteristic storage unit for each time zone. Then, reinforcement learning is performed on the viewing feature data by time zone using the program feature data, and the learning result is registered and updated in the viewing feature storage unit 116 by time zone.

  The viewing feature storage unit 116 for each time zone stores viewing feature data for each time zone indicating the features of the program viewed on the television set at home for each time zone for each day of the week. An example of viewing characteristic data by time zone is shown in FIG. The viewing characteristic data for each time zone includes data such as a keyword of content for each day of the week and each time zone, and a weight corresponding to the keyword. The time zone is set by dividing the day into several slots. In the example of FIG. 7, a day is divided into 8 slots every 3 hours.

  The viewer information estimation unit 117 receives input data indicating the designated day of the week and time from the estimated viewer information acquisition device 12, and views the viewing characteristic data by time zone corresponding to the designated day of the week and time by time zone. The data is read from the feature storage unit 116 and sent to the similarity calculation unit 118. In addition, the viewer information estimation unit 117 transmits the viewing estimation data from the similarity calculation unit 118 to the viewer information acquisition device 12.

  The similarity calculation unit 118 calculates the similarity between the viewing feature data by time zone from the viewer information estimation unit 117 and the family feature data of all family members registered in the family feature database 119, and the calculation result (Viewing estimation degree) is sent to the viewer information estimation unit 117.

  An example of the family feature database 119 is shown in FIG. The family feature data stored in the family feature database 119 includes a keyword indicating the viewing feature of each family member and weight data corresponding thereto.

  Next, operation | movement of the viewer estimation apparatus 10 of this embodiment is demonstrated.

  First, the processing operation in the scene A in which the viewer performs an input operation for viewing the television and the corresponding operation history is stored in the operation history database 112 will be described with reference to the flowchart of FIG.

  When the operation device 11 transmits an operation signal in response to the input operation of the viewer, the operation history generation unit 111 receives this signal and acquires operation content information (step A1). In addition, the operation history generation unit 111 acquires the metadata of the operation target program if it is registered in a database or the like (step A2). Then, operation history data is generated based on the acquired information and registered in the operation history database 112 (step A3). As a result, the operation history of the viewer is accumulated in the operation history database 112.

  Next, the operation history registered in the operation history database 112 is analyzed, and the processing operation in the scene B in which the viewing feature data by time zone is registered in the viewing feature storage unit 116 by time zone is described with reference to the flowchart of FIG. I will explain.

  The operation history determination unit 113 determines whether there is an unprocessed operation history in the operation history database 112 (step B1). Specifically, an unprocessed flag set for each operation history data is referred to, and it is determined whether there is operation history data with an unprocessed flag. If there is no operation history data for which the unprocessed flag is set, this process is terminated (step B1: NO).

  If there is an unprocessed operation history (step B1: YES), one operation history data is read (step B2).

  The operation history determination unit 113 determines whether or not the read operation history data is valid data for estimating the viewer (step B3).

  When it is determined that the operation history data is not valid data (step B3: NO), the operation history determination unit 113 removes the unprocessed flag of the operation history data. Then, the flow returns to step B1.

  When it is determined that the operation history data is valid (step B3: YES), the operation history determination unit 113 passes the operation history data to the viewing feature analysis unit 115.

  For example, when the operation history data read in step B2 is the data of the operation history ID “9828” in FIG. 3, the operation history data includes the content ID “12345678”, and thus is determined as valid data.

  The viewing feature analysis unit 115 reads the program feature data of the viewed program from the program feature database 114 based on the content ID included in the operation history data from the operation history determination unit 113 (step B4).

  Also, the viewing feature analysis unit 115 reads the viewing feature data by time zone corresponding to the operation date and time included in the operation history data from the operation history determination unit 113 from the viewing feature storage unit by time zone 116 (step B5).

  The viewing feature analysis unit 115 uses the program feature data acquired in step B4 to perform reinforcement learning processing on the viewing feature data classified by time period acquired in step B5 (step B6). The viewing feature analysis unit 115 registers the learned result in the viewing feature storage unit 116 by time as new viewing feature data by time zone (step B7). Then, the flow returns to step B1.

  For example, when the operation history data read in step B2 is the data of the operation history ID “9828” in FIG. 3, the viewing feature analysis unit 115 reads the feature data corresponding to the content ID “12345678” in step B4. Further, in step B5, viewing characteristic data classified by time zone corresponding to the operation date and time “December 21, 2009, 13:20” is read. When the day of the week identified from the operation date “December 21, 2009” is “Monday”, the viewing feature analysis unit 115 uses “Monday” as the day of the week in the viewing feature data by time zone of FIG. The viewing characteristic data by time zone of “afternoon (12-15 o'clock)” is read out.

  The learning processing method in step B6 is arbitrary. As illustrated in FIG. 11, the keyword vector of the program feature data and the keyword vector of the viewing feature data by time zone may be added together. Also, for example, when the interval between the date and time of a certain operation and the date and time of the operation immediately after that is very short (predetermined time or less), there is a negative element such as that the viewer does not like the program very much or is selected by mistake. Conceivable. In such a case, the keyword vectors may not be added but may not be learned or negative learning may be performed.

  The execution timing of the process in the scene B described above may be synchronized with the process in the scene A or may be asynchronous.

  Next, the processing operation in the scene C in which the viewer estimation degree is calculated in response to the designation input of the day of the week and the time zone of the viewer estimation target will be described with reference to the flowchart of FIG.

  When the user inputs the day of the week and the time zone for which the viewer is to be estimated to the estimated viewer information acquisition device 12, the estimated viewer information acquisition device 12 sends the information on the day of the week and the time zone specified by the input to the viewer estimation device 10. Transmit (step C1).

  The viewer information estimation unit 117 acquires viewing feature data by time zone corresponding to the designated day of the week and time zone from the viewing feature storage unit by time zone 116 and passes it to the similarity calculation unit 118 (step C2).

  The similarity calculation unit 118 reads the family feature data of all family members from the family feature database 119, calculates the similarity between each family feature data and the viewing feature data classified by time period acquired in step C2, and calculates the calculation result. It passes to the viewer information estimation part 117 (step C3). The similarity calculated here indicates the viewing estimation degree for each family member at the designated time. As a method for calculating the similarity, for example, cosine similarity may be used, or another method capable of calculating the similarity between data may be used.

  The viewer information estimation unit 117 transmits the data of the viewing estimation degree to the estimated viewer information acquisition device 12 (step C4). At this time, viewing characteristic data for each time zone may be transmitted together.

  The calculation of the similarity will be described with reference to FIG. 13 by taking as an example a case where the day of the week “Monday” and the time zone “afternoon (12-15 o'clock)” are designated from the estimated viewer information acquisition apparatus 12.

  The listener information estimation unit 117 acquires the viewing feature data by time zone corresponding to the designated “Monday” and “afternoon (12:00 to 15:00)” from the viewing feature storage unit by time zone 116. Then, the family feature data and similarity of each family member stored in the family feature database 119 are calculated. In the example of FIG. 13, the cosine similarity between the keyword vector of the viewing feature data for each time zone of “Monday” and “Afternoon (12:00 to 15:00)” and the keyword vector of the family feature data of each family is calculated. The calculated similarity (viewing estimate) is 0 for the father and the eldest daughter, 0.91 for the mother, and 0.51 for the eldest son.

  As described above, according to the present embodiment, the viewing feature data corresponding to the designated day of the week and time zone is acquired from the database, and the similarity with the family feature data is calculated for each family member. Therefore, the degree of being considered as a viewer at a specified time can be shown for each family member only with an operation history of the television or the like. Thereby, it is possible to estimate the viewer at an arbitrary time without applying an operation load to the user and with a simple system configuration.

  As a modification of the present embodiment, family feature data may be learned to update the family feature database 119. A block diagram of the viewer estimation device 10 in this case is shown in FIG.

  The viewer estimation device 10 according to the modification of the present embodiment further includes a family feature learning unit 120. In addition, the operation device 11 ′ transmits viewer information for identifying the viewer together with information on the operation content. The method for acquiring the viewer information is arbitrary, and may be performed by a predetermined input operation from the input device, or may be acquired from personal information stored in the mobile phone. The operation history generation unit 111 ′ converts the received viewer information into a viewer ID managed in the family feature database 119, and adds the viewer ID to the operation history data. FIG. 15 shows an example of the operation history database 112 ′ in which operation history data to which the viewer / viewer ID is added is stored.

  The processing operation of the viewer estimation device 10 according to a modification of the present embodiment will be described with reference to the flowchart of FIG. This process is a process in the scene B in which the viewing feature data for each time zone is learned and registered in the viewing feature storage unit for each time zone. In addition, about the operation | movement similar to this embodiment, the same code | symbol is attached | subjected and description is abbreviate | omitted.

  The viewing feature analysis unit 115 ′ determines whether or not the operation history data includes a viewer ID after registering the viewing feature data classified by time period after learning in Step B7 (Step B8).

  If the viewer ID is not included in the operation history (step B8: NO), the flow returns to step B1.

  When the viewer ID is included (step B8: YES), the viewing feature analysis unit 115 ′ sends the operation history data and the corresponding program feature data to the family feature learning unit 120. The family feature learning unit 120 acquires family feature data corresponding to the viewer ID included in the operation history data from the family feature database 119, performs reinforcement learning processing on the family feature data, and registers the learning result in the family feature database 119. (Step B9). The learning processing method may be the same as the processing method of step B6. Thereafter, the flow returns to step B1.

  Thereby, reinforcement characteristic learning of family feature data can be performed, and viewer estimation can be performed following changes in family viewing content.

The case where the viewer estimation apparatus 10 which concerns on embodiment of this invention is implement | achieved with a computer is demonstrated. The computer is a CPU (Central Processing Unit), ROM (Read Only)
A control unit including a memory (RAM) and a random access memory (RAM), and a storage unit including a hard disk device. The control unit of the computer reads and executes an operation program or the like stored in advance in the storage unit, thereby performing an operation history generation unit 111, an operation history determination unit 113, a viewing feature analysis unit 115, and a viewer information estimation unit. 117 and the similarity calculation unit 118 are logically realized. The storage unit includes an operation history database 112, a program feature database 114, an hourly viewing feature storage unit 116, and a family feature database 119.

  The device of the present invention described above can be realized by a computer program as is apparent from the above description, but can also be configured by hardware. Also, only some functions of the above-described embodiments can be realized by a computer program.

  Although the present invention has been described with reference to the preferred embodiments, the present invention is not necessarily limited to the above-described embodiments, and various modifications can be made within the scope of the technical idea.

(Appendix 1)
Time-based viewing features that are analyzed based on the operation history of the viewing device and that store at least one feature related to a program viewed in each time zone and numerical values corresponding to the features are stored. A storage unit;
A family feature storage unit that stores family feature data that quantitatively represents the features of the viewing content for each member of the family;
A similarity calculation unit that extracts viewing feature data by time zone in a specified time zone and calculates the similarity with the family feature data;
A viewer estimation device comprising:

(Appendix 2)
An operation history storage unit for storing operation history data of a viewing device including data of a viewed program and operation date and time;
A program feature storage unit for storing program feature data quantitatively representing the features of each program;
A viewing feature analysis unit that acquires the program feature data corresponding to the viewing program of the operation history data, and learns the viewing feature data classified by time zone corresponding to the operation date and time of the operation history data with the program feature data;
The viewer estimation device according to attachment 1, further comprising:

(Appendix 3)
With respect to the operation history data including the viewer identification information, a family feature learning unit that acquires program feature data corresponding to the operation history data and learns the family feature data corresponding to the viewer identification information with the program feature data Further comprising
The viewer estimation device according to supplementary note 2, characterized by:

(Appendix 4)
The viewing device includes a television and a device connected to the television.
The viewer estimation device according to any one of supplementary notes 1 to 3, wherein

(Appendix 5)
Analyzed on the basis of the operation history of the viewing device, at least one feature relating to the program watched in each time zone, viewing feature data by time zone representing a numerical value corresponding to the feature, and viewing content for each member of the family And family feature data that quantitatively represents the features of
Take out the viewing feature data for each designated time zone and calculate the similarity with the family feature data.
A viewer estimation method characterized by the above.

(Appendix 6)
Create program feature data that quantitatively represents the features of each program,
Storing operation history data of the viewing device including data of the viewed program and operation date and time,
Acquiring the program feature data corresponding to the viewing program of the operation history data, and learning the viewing feature data classified by time zone corresponding to the operation date and time of the operation history data with the program feature data;
The viewer estimation method according to appendix 5, characterized in that

(Appendix 7)
For the operation history data including viewer identification information, obtain program feature data corresponding to the operation history data, and learn the family feature data corresponding to the viewer identification information with the program feature data.
The viewer estimation method according to appendix 5 or 6, wherein the viewer estimation method is characterized.

(Appendix 8)
The viewing device includes a television and a device connected to the television.
The viewer estimation method according to any one of appendices 5 to 7, characterized in that:

(Appendix 9)
Analyzed based on the operation history of the viewing device and at least one feature related to the program viewed in each time zone, viewing feature data by time zone representing a numerical value corresponding to the feature, and viewing content for each family member A computer that uses family feature data that quantitatively represents the features of
A program for realizing a similarity calculation function for extracting viewing feature data by time zone in a specified time zone and calculating a similarity with the family feature data.

(Appendix 10)
The computer further uses operation history data of the viewing apparatus including data of the viewed program and operation date and time, and program feature data that quantitatively represents features of each program,
A viewing feature that acquires the program feature data corresponding to the viewing program of the operation history data to the computer and learns the viewing feature data classified by time zone corresponding to the operation date and time of the operation history data with the program feature data The program according to appendix 9 for further realizing the analysis function.

(Appendix 11)
In the computer,
With respect to the operation history data including the viewer identification information, family feature learning function for acquiring program feature data corresponding to the operation history data and learning the family feature data corresponding to the viewer identification information with the program feature data The program according to appendix 9 or 10 for further realizing the above.

(Appendix 12)
The viewing device includes a television and a device connected to the television.
The program according to any one of appendices 9 to 11, characterized in that:

1 viewing feature data by time zone 2 family feature data 3 similarity calculation unit 10 viewer estimation device 11 operation device 12 estimated viewer information acquisition device 111 operation history generation unit 112 operation history database 113 operation history determination unit 114 program feature database 115 Viewing feature analysis unit 116 Viewing feature storage unit by time zone 117 Viewer information estimation unit 118 Similarity calculation unit 119 Family feature database 120 Family feature learning unit

Claims (8)

Time-based viewing features that are analyzed based on the operation history of the viewing device and that store at least one feature related to a program viewed in each time zone and numerical values corresponding to the features are stored. A storage unit;
A family feature storage unit that stores family feature data that quantitatively represents the features of the viewing content for each member of the family;
A similarity calculation unit that extracts viewing feature data by time zone in a specified time zone and calculates the similarity with the family feature data;
A viewer estimation device comprising: An operation history storage unit for storing operation history data of a viewing device including data of a viewed program and operation date and time;
A program feature storage unit for storing program feature data quantitatively representing the features of each program;
A viewing feature analysis unit that acquires the program feature data corresponding to the viewing program of the operation history data, and learns the viewing feature data classified by time zone corresponding to the operation date and time of the operation history data with the program feature data;
The viewer estimation device according to claim 1, further comprising: With respect to the operation history data including the viewer identification information, a family feature learning unit that acquires program feature data corresponding to the operation history data and learns the family feature data corresponding to the viewer identification information with the program feature data Further comprising
The viewer estimation apparatus according to claim 2, wherein The viewing device includes a television and a device connected to the television.
The viewer estimation device according to claim 1, wherein the viewer estimation device is a viewer. Analyzed on the basis of the operation history of the viewing device, at least one feature relating to the program watched in each time zone, viewing feature data by time zone representing a numerical value corresponding to the feature, and viewing content for each member of the family And family feature data that quantitatively represents the features of
Take out the viewing feature data for each designated time zone and calculate the similarity with the family feature data.
A viewer estimation method characterized by the above. Create program feature data that quantitatively represents the features of each program,
Storing operation history data of the viewing device including data of the viewed program and operation date and time,
Acquiring the program feature data corresponding to the viewing program of the operation history data, and learning the viewing feature data classified by time zone corresponding to the operation date and time of the operation history data with the program feature data;
The viewer estimation method according to claim 5, wherein Analyzed based on the operation history of the viewing device and at least one feature related to the program viewed in each time zone, viewing feature data by time zone representing a numerical value corresponding to the feature, and viewing content for each family member A computer that uses family feature data that quantitatively represents the features of
A program for realizing a similarity calculation function for extracting viewing feature data by time zone in a specified time zone and calculating a similarity with the family feature data. The computer further uses operation history data of the viewing apparatus including data of the viewed program and operation date and time, and program feature data that quantitatively represents features of each program,
A viewing feature that acquires the program feature data corresponding to the viewing program of the operation history data to the computer and learns the viewing feature data classified by time zone corresponding to the operation date and time of the operation history data with the program feature data The program according to claim 7 for further realizing an analysis function.

Images