Interest acquisition device and interest acquisition program

Abstract

To provide an interest acquisition device that enables a robot which views content together with a user to acquire its own interests.SOLUTION: An interest acquisition device 1 comprises: a user interest degree calculation unit 12 which calculates, for each period, an interest degree in a keyword based on a frequency of appearance of the keyword in EPG data of programs viewed by a user of a robot during a predetermined period; a robot interest variation degree calculation unit 14 which calculates an interest degree in the keyword based on a frequency of appearance of the keyword in EPG data of programs viewed by other users during the predetermined period, and calculates a keyword interest variation degree by dividing a difference from a previous period by the number of programs including the keyword; and a robot interest degree calculation unit 15 which calculates an interest degree of the robot in the keyword by accumulating the keyword interest variation degree according to the number of programs including the keyword in programs viewed by the user of the robot.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a technology for controlling the interest of a robot that watches a television program together with a person. [Background technology]

[0002] Research and development of communication robots that can converse with humans has been progressing. For example, Patent Document 1 proposes a technology that extracts keywords from the video, audio, and subtitle data of a television program to generate speech sentences, allowing a robot that watches television with a person to make speech related to the television program, thereby enabling a dialogue between the robot and a person while watching television.

[0003] Patent Document 2 proposes a device that uses a video of a person watching a program content to estimate the degree of interest in the content being watched, based on the person's movements, line of sight, and expressions while watching. Furthermore, Non-Patent Document 1 proposes a device that uses the viewing history of program content (referred to as programs or television programs as appropriate) to extract keywords from the Electronic Programming Guide (EPG) data of programs viewed by a person, and estimates the degree of interest in keywords related to the program based on the maximum TF-IDF (Term Frequency - Inverse Document Frequency) value of each keyword.

[0004] Patent document 3 proposes a pet robot that uses the user's behavioral history to calculate the degree of intimacy and determine how to approach the user, thereby autonomously expressing emotions and performing actions such as showing shyness, thereby creating an effect of attachment and providing comfort to the user. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2019-185400 A [Patent Document 2] Japanese Patent Application Laid-Open No. 2013-109537 [Patent Document 3] Japanese Patent Application Publication No. 2019-5591 [Non-patent literature]

[0006] [Non-Patent Document 1] Makoto Okuda et al., "Estimating TV Viewer Interests by Analyzing Electronic Program Guides," ITE Technical Report, Vol. 46, No. 13, 2022. Summary of the Invention [Problem to be solved by the invention]

[0007] However, even if conventional techniques attempt to estimate the user's interests, they do not allow the robot that watches television with the user to have its own interests. When a robot speaks based on the user's interests while watching TV with the user, it has the effect of making the user feel a sense of enjoyment and attachment to the robot. On the other hand, speaking only based on the user's interests is not expected to have the effect of giving the user new discoveries. Even when people watch TV together, they can not only empathize with each other's interests, but also discover new interests by having a conversation about the other person's interests that are different from their own.

[0008] For this reason, it is desirable for a robot that watches television with a person to acquire the user's interests while also having its own interests that differ from the user's. Also, just as people's interests change as they influence each other's interests, it is desirable for the robot's interests to change as it watches television with the user. An object of the present invention is to provide an interest acquisition device and an interest acquisition program that allow a robot that watches content together with a user to acquire its own interests while being influenced by the user's interests. [Means for solving the problem]

[0009] The interest acquisition device of the present invention comprises: a program information acquisition unit that acquires program guide data for programs viewed by a user of a robot and other users different from the user during a predetermined period; a keyword extraction unit that extracts keywords from the program guide data; a user interest calculation unit that uses the program guide data for programs viewed by the user of the robot to calculate, for each period, a user interest level for the keyword based on the frequency of appearance of the keyword in each of the program guide data, and the number of programs including the keyword; a robot interest change level calculation unit that uses the program guide data for programs viewed by the other users to calculate, for each period, a user interest level for the keyword of the other users based on the frequency of appearance of the keyword in each of the program guide data, and the number of programs including the keyword, and calculates a keyword interest change level by prorating the difference in interest level for the keyword from the previous period by the number of programs including the keyword; and a robot interest level calculation unit that calculates a value obtained by accumulating the keyword interest change level according to the number of programs including the keyword calculated by the user interest level calculation unit, as the robot interest level for the keyword.

[0010] The interest acquisition device may further include a program genre extraction unit that extracts a program genre from the program guide data acquired by the program information acquisition unit, and the user interest calculation unit may use the program guide data of the programs viewed by the user of the robot to calculate, for each period, the user's interest in the program genre based on the frequency of appearance of the program genre in each of the program guide data, and the number of programs of the program genre; the robot interest change calculation unit may use the program guide data of the programs viewed by the other users to calculate, for each period, the other users' interest in the program genre based on the frequency of appearance of the program genre in each of the program guide data, and the number of programs of the program genre, and calculate a program genre interest change degree by proportionally dividing the difference in the interest in the program genre from the previous period by the number of programs of the program genre; and the robot interest degree calculation unit may calculate, as the robot's interest in the program genre, a value obtained by accumulating the program genre interest change degrees according to the number of programs of the program genre calculated by the user interest calculation unit.

[0011] The user interest calculation unit may calculate the maximum TF-IDF value of each keyword when program guide data for each program viewed by the user of the robot is considered as a single document as the user's interest in the keyword, and calculate the frequency of appearance of the program genre as the user's interest in that program genre, and the robot interest change calculation unit may calculate the maximum TF-IDF value of each keyword when program guide data for each program viewed by the other user is considered as a single document as the other user's interest in the keyword, and calculate the frequency of appearance of the program genre as the other user's interest in that program genre.

[0012] The interest acquisition device may include a program interest calculation unit that calculates the user's interest level and the robot's interest level in the program to be viewed based on the user's interest level and the robot's interest level for each of keywords and program genres included in the program guide data of the program to be viewed.

[0013] The program interest level calculation unit may select a program the user wishes to view based on a product of the user's interest level in the candidate programs and the robot's interest level.

[0014] An interest acquisition program according to the present invention is for causing a computer to function as the interest acquisition device. [Effects of the Invention]

[0015] According to the present invention, a robot that watches content together with a user is influenced by the user's interests and acquires its own interests. [Brief explanation of the drawings]

[0016] [Figure 1] 1 is a block diagram showing a functional configuration of an interest acquisition device according to an embodiment. [Figure 2] 10 is a diagram illustrating processing details of a user viewing history acquisition unit in the embodiment. FIG. [Figure 3] 10 is a flowchart showing a procedure for calculating a user's interest level in an embodiment. [Figure 4] 10A and 10B are diagrams illustrating processing details performed by a keyword interest degree change unit in the embodiment. [Figure 5] 10A and 10B are diagrams illustrating processing contents performed by a keyword interest level unit in the embodiment. [Figure 6] 10 is a flowchart showing a procedure for calculating the interest level of a robot in the embodiment. [Figure 7] 10 is a diagram illustrating processing performed by a user degree of interest in programs unit in the embodiment. FIG. [Figure 8] FIG. 10 is a diagram illustrating the processing contents of the user / robot program interest level unit in the embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0017] An example of an embodiment of the present invention will be described below. The interest acquisition device of this embodiment is a robot that watches content such as television programs together with a person, and acquires the robot's own interest and utilizes it in communication with the user.

[0018] FIG. 1 is a block diagram showing the functional configuration of an interest acquisition device 1 according to this embodiment. The interest acquisition device 1 is an information processing device (computer) equipped with a control unit and a memory unit, and is incorporated into a robot or communicatively connected to the robot to acquire the robot's own interest while outputting program information selected based on the interests of both the user and the robot.

[0019] Data included in an electronic program guide (EPG) for a program currently being played is input from a television (television receiver) to the interest acquisition device 1. Furthermore, the input data is stored in a predetermined database as a viewing history (individual record) for each user. The interest acquisition device 1 is also connected via a network to a database (EPG information DB) that stores EPG data and is provided in an external server or the like, and receives data stored in this EPG information DB.

[0020] The control unit operates as the following functional units by reading and executing software (interest acquisition program) stored in the storage unit. In addition to the interest acquisition program, the storage unit also stores various databases (DB). Specifically, the control unit includes a user viewing history acquisition unit 11, a user interest calculation unit 12, an individual viewing history acquisition unit 13, a robot interest change calculation unit 14, a robot interest calculation unit 15, and a program interest calculation unit 16. The storage unit also includes a user interest level DB 21, a robot interest level DB 22, and a program interest level DB 23.

[0021] The user viewing history acquisition unit 11 includes an EPG information acquisition unit 111 (program information acquisition unit), a keyword extraction unit 112, and a program genre extraction unit 113, and acquires history information about television programs viewed by the user.

[0022] The EPG information acquisition unit 111 acquires EPG data of television programs viewed by the user of the robot directly from the television or from a predetermined server via a network.

[0023] The keyword extraction unit 112 extracts keywords from the EPG data acquired by the EPG information acquisition unit 111 . Here, the keywords are words such as proper nouns contained in the EPG data, and may be extracted from a dictionary prepared in advance.

[0024] The program genre extraction unit 113 acquires, from the EPG data acquired by the EPG information acquisition unit 111, genre information (program genre) of the program viewed by the user. Here, the program genre is one or more categories set for each program according to its content, and for example, a medium category (sub-genre) in the EPG data may be extracted.

[0025] FIG. 2 is a diagram illustrating the processing contents of the user viewing history acquisition unit 11 in this embodiment. The EPG data includes time information for each program, a code for classifying the content of the program (major category, medium category, etc.), and the program title or introductory text.

[0026] The keyword extraction unit 112 extracts a plurality of keywords such as proper nouns from the text data. The program genre extraction unit 113 extracts program genres such as "0x4f" (variety / other) and "0x01" (weather) from the EPG data. Note that, for example, "4" in "0x4f" is a code indicating a major category and "f" is a code indicating a medium category.

[0027] The user interest calculation unit 12 includes a keyword interest unit 121 and a program genre interest unit 122, and calculates the user's interest in each of the keywords and program genres related to television programs viewed by the user during a predetermined period.

[0028] The keyword interest level unit 121 calculates the user's interest level (user interest level) for each of all keywords extracted by the keyword extraction unit 112. Specifically, the keyword interest level unit 121 organizes information on the EPG data of each program that the user v has viewed in a predetermined period (for example, one month) into one document d. v1 ,d v2 ,…,d vm and the keyword w i Number of documents containing |{d:d∋w i}|(number of programs watched), and document d vk Keyword in w i n occurrences of w_i,d_vk Then, document d vk Keyword w in (k=1,2,…,m) i The TF (Term Frequency) value (tf w_i,d_vk ), and the keyword w i IDF (Inverse Document Frequency) value of (idf w_i ) is calculated as follows, and the maximum TF-IDF value is calculated as the interest level (int w_i,v ) Note that w_i, d_vi are the w in the following formula. i ,d vk In the following, subscripts may be expressed using "_".

number

[0029] The program genre interest level unit 122 calculates the user's interest level (user interest level) for each of all program genres extracted by the program genre extraction unit 113. Specifically, the program genre interest level unit 122 counts the number of programs viewed for each program genre in a predetermined period and the sum of the number of programs viewed for all program genres, and calculates a TF value (tf sg_i,v ) to user v's program genre sg i The interest level for is calculated as follows:

number

[0030] The program genre interest level unit 122 stores the calculated interest level for each of all program genres in the user interest level DB 21. The number of programs watched in the program genre used to calculate the interest level is used by the robot interest level calculation unit 15 to calculate the robot's interest level in the program genre.

[0031] FIG. 3 is a flowchart showing the procedure for calculating the user's interest level by the user viewing history acquisition unit 11 and the user interest level calculation unit 12 in this embodiment.

[0032] In step S1, the EPG information acquisition unit 111 acquires EPG data of a television program viewed by the user of the robot.

[0033] In step S2, the keyword extraction unit 112 extracts keywords such as proper nouns from the EPG data acquired in step S1. In step S3, the keyword interest level unit 121 calculates the user's interest level for each of the keywords extracted in step S2.

[0034] In step S4, the program genre extraction unit 113 extracts the program genre from the EPG data acquired in step S1. In step S5, the program genre interest level unit 122 calculates the user's interest level in each of the program genres extracted in step S4.

[0035] In step S6, the user interest degree calculation unit 12 (the keyword interest degree unit 121 and the program genre interest degree unit 122) stores the interest degrees calculated in steps S3 and S5 in the user interest degree DB 21.

[0036] The individual viewing history acquisition unit 13 includes an EPG information acquisition unit 131, a keyword extraction unit 132, and a program genre extraction unit 133, and acquires history information about programs viewed by users other than the user of the robot. Here, the viewing histories of multiple users who own robots of the same model, including the target robot, are stored on a predetermined server, and are shared by multiple robots of the same model on a network. These viewing histories may be, for example, obtained by the EPG information acquisition unit 111 after acquiring EPG data and then accessing the server and storing them. As a result, each interest acquisition device 1 corresponding to each robot can designate a specific person other than its own user as an "other user," acquire that person's viewing history from the server, and use that person's viewing history as information on a personality different from that of the robot's user. Since it is difficult to extract common keywords or program genres from this "other user" if their hobbies and preferences differ from those of the robot's user, it is preferable to select individuals who match the robot's user's age, gender, etc., rather than selecting them completely randomly.

[0037] The EPG information acquisition unit 131 acquires EPG data from the viewing history (individual records) of other users from a predetermined server via a network. The keyword extraction unit 132 and the program genre extraction unit 133 have the same functions as the keyword extraction unit 112 and the program genre extraction unit 113 of the user viewing history acquisition unit 11, respectively, and extract keywords and program genres from the EPG data acquired by the EPG information acquisition unit 131.

[0038] The robot interest change degree calculation unit 14 includes a keyword interest change degree unit 141 and a program genre interest change degree unit 142, and calculates the change degree of interest of other users in the keywords and program genres extracted by the individual viewing history acquisition unit 13.

[0039] The keyword interest degree change unit 141 has the same function as the keyword interest degree unit 121 of the user interest degree calculation unit 12, and calculates the degree of interest and the number of programs watched by other users for each of all keywords extracted by the individual viewing history acquisition unit 13 for the same period, for example, monthly, as the user interest degree calculation unit 12. Furthermore, the keyword interest degree change unit 141 uses the degree of interest in the keyword and the number of programs watched to proportionally divide the difference in the degree of interest in the keyword in a certain period from the previous period by the number of programs watched within the period for that keyword, and calculates a keyword interest degree change according to the number of programs watched for the keyword. Specifically, the keyword for the first month (m1) i Change in interest in (inc w_i,m_1 ) is the interest level (int w_i,v,m_1 ) and the number of programs viewed (c w_i,m_1 ) and the change in interest from the second month (m2) onwards is calculated as follows, using the increase in interest from the previous month and the number of programs watched in the current month.

number

[0040] This keyword interest change degree is an increase or decrease in interest degree that is added up every time the robot user watches a program whose EPG data includes the relevant keyword.

[0041] FIG. 4 is a diagram illustrating the processing performed by the keyword interest degree variation unit 141 in this embodiment. For example, the interest level for the keyword "baseball" was 0.08 in the first month and 0.06 in the second month, and the number of programs watched containing the keyword was 20 and 18, respectively. In this case, the change in interest up to the 20th program viewed in the first month is 0.08 / 20 = 0.004. The change in interest from the 21st program to the 38th program viewed is (0.06 - 0.08) / 18 = -0.001.

[0042] The program genre interest degree change unit 142 has the same function as the program genre interest degree unit 122 of the user interest degree calculation unit 12, and calculates the degree of interest and the number of programs watched by other users for each of all program genres extracted by the individual viewing history acquisition unit 13 for the same period as the user interest degree calculation unit 12. Furthermore, the program genre interest degree change unit 142 calculates the interest degree according to the number of programs watched for each program genre, similar to the keyword interest degree change unit 141, using the interest degree for each program genre and the number of programs watched. The interest change rate is calculated for all months. For example, if the individual record shows viewing history from January 2016 to December 2018, the interest change rate for 36 months is calculated.

[0043] The robot interest level calculation unit 15 includes a keyword interest level unit 151 and a program genre interest level unit 152, and calculates the robot's interest level (robot interest level) in the keyword and program genre using the number of programs watched by the robot's user and the interest change rate of other users for each keyword and program genre.

[0044] The keyword interest degree unit 151 integrates the keyword interest change degree calculated by the keyword interest change degree unit 141 according to the number of programs in which the keyword appeared (number of programs viewed) c counted by the keyword interest degree unit 121 as follows: i Interest level (int w_i,c_m ) where inc w_i,c indicates the degree of interest change.

number

[0045] FIG. 5 is a diagram illustrating the processing contents of the keyword interest level unit 151 in this embodiment. For example, if the change in interest level for the keyword "baseball" is "0.004" up to the 20th time and "-0.001" from the 21st time onwards, and if the number of programs in which "baseball" appeared (number of programs watched) watched by the robot user was 21, then the degree of interest in "baseball" would be 0.079, calculated by accumulating 0.004 20 times and -0.001 once.

[0046] In addition, if the interest change degree has not been calculated up to the number of programs viewed based on the viewing history of the robot user, i.e., if the number of programs viewed by other users using the same keyword is small, the interest change degree for the relevant episode will be set to 0, and the interest degree will not change even if the keyword appears again. The keyword interest level unit 151 stores the calculated interest level of the robot for each of all keywords in the robot interest level DB 22.

[0047] Similar to the keyword interest level unit 151, the program genre interest level unit 152 adds up the program genre interest change level calculated by the program genre interest change level unit 142 according to the number of programs of the program genre (number of programs viewed) among the programs viewed by the robot user, as counted by the program genre interest level unit 122, and calculates it as the robot's interest level in the program genre. The program genre interest level unit 152 stores the calculated robot interest levels for each of all program genres in the robot interest level DB 22.

[0048] FIG. 6 is a flowchart showing the procedure for calculating the degree of interest in the robot by the individual viewing history acquisition unit 13, the robot interest change degree calculation unit 14, and the robot interest degree calculation unit 15 in this embodiment.

[0049] In step S11, the EPG information acquisition unit 131 acquires, from the individual records, EPG data of television programs viewed by users other than the user of the robot.

[0050] In step S12, the keyword extraction unit 132 extracts keywords such as proper nouns from the EPG data acquired in step S11. In step S13, the keyword interest degree change unit 141 calculates the interest degree change for each of the keywords extracted in step S12. In step S14, the keyword interest level unit 151 calculates the robot's interest level in the keyword by accumulating the change in interest level in the keyword calculated in step S13 according to the number of programs watched for that keyword counted by the keyword interest level unit 121.

[0051] In step S15, the program genre extraction unit 133 extracts the program genre from the EPG data acquired in step S11. In step S16, the program genre interest change degree unit 142 calculates the interest change degree for each of the program genres extracted in step S15. In step S17, the program genre interest level unit 152 calculates the robot's interest level in the program genre by accumulating the change in interest level in the program genre calculated in step S16 according to the number of programs watched in that program genre counted by the program genre interest level unit 122.

[0052] In step S18, the robot interest level calculation unit 15 (the keyword interest level unit 151 and the program genre interest level unit 152) stores the interest levels calculated in steps S14 and S17 in the robot interest level DB 22.

[0053] The program interest calculation unit 16 includes a program information acquisition unit 161, a user program interest unit 162, a robot program interest unit 163, and a user / robot program interest unit 164, and calculates the interest levels of both the user and the robot for each program scheduled to be broadcast as a candidate for viewing. The program information acquisition unit 161 acquires EPG data of television programs scheduled to be broadcast in the near future (for example, within one or several days, or within one week) directly from the television or from a predetermined server via a network. The program information acquisition unit 161 further extracts keywords and program genres for each program from the acquired EPG data.

[0054] The user interest level in programs unit 162 calculates the user's interest level in each program in the EPG data acquired by the program information acquisition unit 161, using the user's interest level in each keyword and program genre in the user interest level DB 21. Specifically, the user interest level unit 162 calculates the highest interest level for each keyword and program genre for each program, and multiplies the result as follows to obtain the interest level for program a (int a )

number

[0055] FIG. 7 is a diagram illustrating the processing performed by the user degree of interest in programs unit 162 in this embodiment. From Program 1 and Program 2 in the EPG data, one or more program genres and keywords (two for each in FIG. 7) are extracted by the program information acquisition unit 161. The user program interest level unit 162 extracts the user's interest level for each of these program genres and keywords from the user interest level DB 21.

[0056] For example, for Program 1, the user's interest level for each program genre is highest for Program Genre 1 at 0.06. Also, the user's interest level for each keyword is highest for Keyword 3 at 0.03. Therefore, the user's interest level for Program 1 is 0.06 x 0.03 = 0.0018. Similarly, the user's interest level in Program 2 is highest for Program Genre 2 and Keyword 7, so it is 0.02×0.04=0.0008.

[0057] The robot program interest level unit 163 calculates the robot's interest level for each program in the EPG data acquired by the program information acquisition unit 161 in the same way as the user program interest level unit 162, using the robot's interest level for each keyword and program genre in the robot interest level DB22.

[0058] The user / robot degree of interest in programs unit 164 calculates the user / robot degree of interest in programs using the degrees of interest in each program calculated by the user / robot degree of interest in programs unit 162 and the robot / robot degree of interest in programs unit 163, respectively. Specifically, the user / robot program interest level unit 164 calculates the user's interest level (int k,h ) and the robot's interest level (int k,r ) and multiply it by the value (int k,h,r ) is stored as the user robot program interest level for each program in the program interest level DB 23. Furthermore, the user robot program interest level unit 164 selects the program with the highest user robot program interest level from the programs in the EPG data acquired by the program information acquisition unit 161, and presents it to the user.

number

[0059] FIG. 8 is a diagram illustrating the processing performed by the user / robot program interest level unit 164 in this embodiment. In this example, the user's interest level and the robot's interest level are obtained for programs 1, 2, 3, etc. The value obtained by multiplying these two interest levels for each program (user-robot interest level) is the highest for program 1. Therefore, the user / robot program interest level unit 164 selects program 1 as a program that is interesting to both the user and the robot. The user / robot interest level unit 164 may select a predetermined number (for example, three) of programs in descending order of the user / robot interest level in programs, or may select programs that are equal to or greater than a preset threshold value.

[0060] In this case, even if either the user or the robot has a high level of interest, the program is unlikely to be selected if the other has a low level of interest, and a program that both users and the robot are somewhat interested in is likely to be selected. Therefore, a program that both the user and the robot are interested in, and at the same time, that each has different interests, is likely to be selected.

[0061] According to this embodiment, the interest acquisition device 1 uses the viewing history (individual records) of other users in addition to the user's viewing history when calculating the robot's interest level, so that the robot's interest is influenced by the user's interest while also acquiring its own unique interest. That is, the interest acquisition device 1 calculates the interest change degree of other users for each keyword and program genre from the individual records. Furthermore, the interest acquisition device 1 counts the number of programs viewed for each keyword and each program genre from the viewing history together with the user, i.e., the user's viewing history, and adds up the interest change degrees for the number of programs viewed, thereby making the robot acquire an interest different from that of the user.

[0062] This allows the robot to acquire its own interests while being influenced by the user's interests (viewing history). As a result, for example, from a week's worth of EPG data, the robot can find programs to watch together with the user. In this embodiment, it is also possible to select a program to watch based only on the degree of interest in keywords, without using the program genre. In this case, in Figure 7, the user's degree of interest in program 1 is 0.03 for keyword 3, and the user's degree of interest in program 2 is 0.04 for keyword 7. However, by also taking the degree of interest in program genre into consideration, it is possible to prevent inconveniences such as programs with short program introductions in the EPG being difficult to select.

[0063] This embodiment is intended to help people discover new programs that they are interested in. To achieve this, a robot is utilized that watches content such as television programs together with people. For example, when two people watch TV together, each person has different interests, and as a result, the person is influenced by the other person's interests and discovers new programs that interest them. In order for a robot to have this same influence on people, the interest acquisition device 1 is configured so that the robot is influenced by the user's interests and develops its own unique interests.

[0064] The interest acquisition device 1 also calculates the interest levels of the robot and the user in keywords and program genres, and selects from the EPG data programs that contain keywords and program genres that are of high interest to both users. The robot notifies the user of the selected programs by speech (e.g., "Would you like to watch program A together, which is about to start?") or by movement. As a result, by watching a program that both users are interested in, rather than a program that only one user is interested in, their interests can influence each other, leading to new discoveries.

[0065] In this way, the interest acquisition device 1 selects a program based on the interests of both the robot and the user, and the robot recommends the program, thereby enabling the user to discover a new interest. In addition, since the robot acquires its own interests while being influenced by the interests of the user, robots with different interests are created for different users, which can give people the effect of becoming attached to them and having fun. Furthermore, by continuing to watch TV with the user, the robot's interests in programs change, which has the effect of eliminating the user's boredom with the robot.

[0066] Although the embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments. Furthermore, the effects described in the above-described embodiments are merely a list of the most preferable effects resulting from the present invention, and the effects of the present invention are not limited to those described in the embodiments.

[0067] In this embodiment, the configuration and operation of the interest acquisition device 1 have been mainly described, but the present invention is not limited to this and may be configured as a method or program that includes each component and is used to acquire the interest of users and robots.

[0068] Furthermore, the functions of the interest acquisition device 1 may be realized by recording a program for realizing the functions of the interest acquisition device 1 on a computer-readable recording medium, and reading and executing the program recorded on the recording medium into a computer system.

[0069] The term "computer system" here includes hardware such as the OS and peripheral devices. Additionally, "computer-readable recording media" refers to portable media such as flexible disks, optical magnetic disks, ROMs, and CD-ROMs, as well as storage devices such as hard disks built into computer systems.

[0070] Furthermore, the term "computer-readable recording medium" may include a medium that dynamically stores a program for a short period of time, such as a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line, or a medium that stores a program for a fixed period of time, such as a volatile memory within a computer system that serves as a server or client in such a case. Furthermore, the program may be one that realizes part of the above-mentioned functions, or may be one that can realize the above-mentioned functions in combination with a program already recorded in the computer system. [Explanation of symbols]

[0071] 1. Interest Acquisition Device 11 User viewing history acquisition unit 12 User Interest Calculation Unit 13 Individual viewing history acquisition unit 14 Robot interest change calculation unit 15 Robot Interest Calculation Unit 16 Program Interest Calculation Unit 21 User Interest DB 22 Robot Interest DB 23 Program Interest DB 111 EPG information acquisition section 112 Keyword Extraction Unit 113 Program Genre Extraction Unit 121 Keyword Interest Section 122 Program Genre Interest Level Section 131 EPG information acquisition section 132 Keyword Extraction Unit 133 Program Genre Extraction Unit 141 Keyword Interest Changes 142 Program Genre Interest Changes Section 151 Keyword Interest Section 152 Program Genre Interest Level Section 161 Program Information Acquisition Unit 162 User Program Interest Level Section 163 Robot Program Interest Level Section 164 User Robot Program Interest Level Section

Claims

1. a program information acquisition unit that acquires program guide data of programs viewed by a user of the robot and other users different from the user during a predetermined period; a keyword extraction unit that extracts keywords from the program guide data; a user interest calculation unit that uses the program guide data of the programs viewed by the user of the robot to calculate, for each period, a user interest in the keyword based on the frequency of appearance of the keyword in each of the program guide data, and the number of programs including the keyword; a robot interest change degree calculation unit that calculates, for each period, the interest level of the other user in the keyword based on the frequency of appearance of the keyword in each of the program guide data of the programs viewed by the other user, and the number of programs including the keyword, and calculates a keyword interest change degree by proportionally dividing the difference in interest level in the keyword from the previous period by the number of programs including the keyword; An interest acquisition device comprising: a robot interest calculation unit that calculates the robot interest level for the keyword by accumulating the keyword interest change level according to the number of programs containing the keyword calculated by the user interest calculation unit.

2. a program genre extraction unit that extracts a program genre from the program guide data acquired by the program information acquisition unit; the user interest calculation unit uses the program guide data of the programs viewed by the user of the robot to calculate, for each period, the user interest in the program genre based on the frequency of appearance of the program genre in each of the program guide data, and the number of programs of the program genre; the robot interest change degree calculation unit uses the program guide data of the programs watched by the other users to calculate, for each period, the degree of interest of the other users in the program genre based on the frequency of appearance of the program genre in each of the program guide data, and the number of programs in the program genre, and calculates a program genre interest change degree by proportionally dividing the difference in the degree of interest in the program genre from the previous period by the number of programs in the program genre; The interest acquisition device according to claim 1, wherein the robot interest calculation unit calculates the robot interest level for the program genre by accumulating the program genre interest change level according to the number of programs in the program genre calculated by the user interest calculation unit.

3. the user interest calculation unit calculates the maximum TF-IDF value of each keyword when program guide data for each program viewed by the user of the robot is regarded as one document as the user interest in the keyword, and calculates the frequency of appearance of the program genre as the user interest in the program genre; The interest acquisition device described in claim 2, wherein the robot interest change degree calculation unit calculates the maximum TF-IDF value of each keyword when the program guide data for each program watched by the other user is considered as a single document as the other user's interest in the keyword, and calculates the frequency of appearance of the program genre as the other user's interest in that program genre.

4. The interest acquisition device according to claim 2, further comprising a program interest calculation unit that calculates the user's interest level and the robot's interest level in the candidate program based on the user's interest level and the robot's interest level for each of the keywords and program genres included in the program guide data for the candidate program.

5. The interest acquisition device according to claim 4 , wherein the program interest calculation unit selects a program the user wishes to view based on a product of the user's interest in the candidate programs and the robot's interest in the programs.

6. An interest acquisition program for causing a computer to function as the interest acquisition device according to any one of claims 1 to 5.

Images