JP2010002944A - Content recommendation device and method thereof - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To recommend proper content even when the storage of user taste information is not sufficient. <P>SOLUTION: An unspecified large number of content keyword information is stored in a server, and content being unspecified large number of taste information and the ranking data of keywords are created. The mixing rate of user side tastes and the unspecified large number of tastes is designated to determine which of those tastes should be preferentially recommended. When the user sets "his or her request to preferentially recommend the unspecified large number of tastes", the content which is not included in the taste ranking data of the user, and which is frequently listened by the other users is recommended based on the unspecified large number of taste information. Also, the user is made to visually view the genre of his or her current tastes, and when the user intuitively sets the significant items of recommendation, the content is mixed according to the tastes of the user. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a signal processing device, a signal processing method, and a program equipped with an Internet browsing function.

  When browsing and creating digital content, there is a system as shown in Patent Document 1 as a content recommendation system that matches user preferences. In this system, it is possible for the user to determine which of the user's preference and topicality information is given priority, and content can be created based on the acquired priority information. Recommendations by genre cannot be made, and there is no mechanism for automatically adjusting the ratio between user preferences and unspecified many preferences based on a specific rule.

JP2003-157278

Problems to be solved by the present invention are as follows.
1) When user operation information cannot be acquired in the initial state, content cannot be recommended. 2) When the recommended content is biased from the beginning even though there is not enough information about the user, user preference is further improved. There is a possibility that the recommended function will have an unfavorable effect due to the biased guidance.
3) If the user's preference changes, it may not be possible to follow this, and therefore content corresponding to the changed preference may be hardly recommended.
4) If only the content that matches the user's preference is recommended, the content is biased toward a specific genre, and the opportunity to view content of other genres is extremely reduced.

In the present invention, the following means are used to solve the problem.
-Unspecified number of content keyword information is stored in the server, and content and keyword ranking data (number of viewing times, number of reservations, number of searches is converted into points) that are unspecified number of preference information.
Specifying the user's preference, giving preference to an unspecified number of preferences, or specifying the mixing ratio.
・ If the user has set "I want you to give preference to an unspecified number of preferences", it is content that is not included in the user's preference ranking data, and other information based on unspecified number of preference information Recommend what most users watch.
-Visually show the current user's favorite genre as a graph, and mix content of different genres according to user preference. The mixing ratio is intuitively determined according to the user's preference.

  In the present invention, even when the user's preference information cannot be acquired in the initial state (when there are few keywords), it can be supplemented and recommended by a large number of unspecified preference information.

  The present invention can be applied to signal processing apparatuses such as a PC, a TV, a recorder, and a mobile phone equipped with an Internet browsing function or a TV recording function.

  FIG. 1 shows an example of a method for adjusting the information amount of user preference and server preference. The user preference and server preference setting bars are adjusted when user preference information (user viewing data) can be collected to some extent and when user preference information is sufficiently collected. At this time, a threshold value is determined according to a mathematical formula, the threshold value is adjusted, and the information amount of user preference and server preference is adjusted. Note that the preference information may be referred to as preference data, viewing data, viewing information, operation data, and operation information.

  101 and 102 are slide positions, 101 is a slide position when the user preference and the server preference are approximately half, and 102 is a slide position in a state where user preferences are sufficiently collected. Reference numeral 103 denotes a setting bar.

FIG. 1A shows a case where user operation information can be collected to some extent, and the position of the slide 101 is the result of the mathematical formula and the threshold value so that the user-side preference and the server-side preference are halved. Adjusted based on. For example, if data mining is used for analyzing user preferences, whether or not user operation information is sufficiently collected is determined based on the certainty factor. Confidence is the rate at which a certain event occurs when a certain condition occurs. The certainty factor is expressed by the following equation.
Confidence level (A → B) = Number of transactions in which both A and B appear / Number of transactions including A In addition, the threshold for determining that user operation information has been sufficiently collected is set to 80% of the total amount of data If you keep
Expression (1) User operation history information / User operation history information (capacity limit) × 100 ≧ threshold (80%)
If the condition of the expression (1) is satisfied, it may be determined that the user's operation history information is sufficiently collected.

  In FIG. 1B, the user's operation information is sufficiently collected, and the position of the slide 102 is adjusted so that the user's preference can be taken into consideration.

  If the user's operation information is not collected enough, the slide position cannot be moved, and if the operation information accumulates to some extent, the slide position can be moved or the user operation information is collected. There is also a method in which the length of the slide position grows with time. Further, since it is not biased only to the user's preference, it should not be completely biased to the user preference and the server preference (the user preference or the server preference becomes 100%).

  Thereby, it is possible to make a recommendation even if the user's operation information is insufficient. Further, a method in which the position of the slide is adjusted not only with the amount of user operation information but also with the passage of time is also conceivable.

  If user preference information is not collected at all (for example, immediately after purchasing a TV), the server preference is set to 100%, and the user moves the slide bar toward the terminal preference as user preference data gathers. it can. In the expression (1), if the value of user operation history information / user operation history information (capacity limit) × 100 is 0%, the slide bar in FIG. Is 100%, the slide bar in FIG.

FIG. 2 is an example of a network configuration.
201 is a PC (client) of other user 1, 202 is a TV of other user 2, 203 is the Internet, 204 is a server, 205 is a user's PC (client), and 206 is a preference analysis result of user 1 (aaa, xxx, yyy) , Zzz are program names), 207 is a preference analysis result of user 2, 208 is a preference analysis result of each PC (client) and TV of 201, 202, and 203, 209 is a preference analysis result of a user, and 210 is a preference analysis result of 208 This is a result (logical product) of the preference analysis result and the preference analysis result of 209.

  Viewing / search / operation histories of the PC (client) 201 of the other user 1, the TV 202 of the other user 2, and the user's PC (client) 205 are collected in one place of the server 204. The server 204 generates a ranking on the server based on such information.

  For example, when a ranking information display command is issued from the PC (client) 201 of the other user 1, the PC (client) 201 of the other user 1 sends a viewing / search / operation history, an analysis result request, and a mixing ratio to the server 204. Send. Upon receiving this request, the server 204 analyzes the preference and transmits the analysis result to the PC (client) 201 of the other user 1. The PC (client) 201 of the other user 1 who has received the analysis result displays the result.

  The preference analysis result 206 of the other user 1 is the PC (client) 201 of the other user 1, the preference analysis result 207 of the other user 2 is the TV 202 of the other user 2, each PC (client), and the preference analysis result 208 of the TV. The preference analysis result 209 of the server 204 and the user's PC (client) is stored in the user's PC (client) 205. The preference analysis result 210 is stored in the user's PC (client) 205 by the user's PC (client) 205 receiving information from the server 204.

  As a result, it is possible to mix user preferences and other user preferences to prevent bias in user preferences, and to find content that is unknown to the user and popular with other users (topical) It becomes. In the future, the user's PC 201, 205, TV 202, etc. may be referred to as a user terminal (or simply “user”).

FIG. 3 is an example of a hardware configuration diagram of the user PC (client) 205 in FIG. 2 and the server 204 in FIG.
301 is a CPU, 302 is a TV (content) viewing application, 303 is a driver, 304 is interface hardware, 305 is a memory, 306 is a program, 307 is a display device, 309 is a mouse, 310 is a keyboard, 308 is an input device, Reference numeral 311 represents the Internet, and 312 represents an auxiliary storage device.

  313 is a server CPU, 314 is a server application, 315 is a server driver, 316 is a server interface and hardware, 317 is a server memory, 318 is a server program, 319 is a server auxiliary storage device, and 201 is a network Another PC 202 connected is another TV connected to the network.

  When the user activates the TV (content) viewing application 302 with the input device 308, the program 306 is written in the memory 305, and the program is executed. The user selects the ratio of the user-side preference and the server-side preference from the input device 308. The selected data is stored in the auxiliary storage device 312. When the user displays the user preference ranking result of the TV (content) viewing application 302 from the input device 308, the mixed ratio of the user preference and the server-side preference stored in the auxiliary storage device 312 and the analysis request are transmitted via the Internet 311. Send.

  The server side 204 similarly obtains information from other PCs 201 and other TVs 202, creates ranking information for the entire user, and stores it in the auxiliary storage device 319. When an analysis request is received from the user 205, the entire result including the preference information of the user 205 and the preference analysis results of the other PC 201 and the other TV 202 is analyzed, and the information is transmitted to the user side. The PC on the user side 205 displays the result sent from the server side 204 on the display device 307.

  In addition to the mouse 309 and the keyboard 310, the input device 308 may use a remote controller, a joystick, a game controller, or the like. As the display device 307, a CRT display, a liquid crystal panel, a plasma panel, or the like may be used.

  The program 306 includes an individual preference ranking generation program that generates individual preference ranking information indicating user preferences from user viewing information, a reception program that receives server-side preference ranking information sent from the server, and a user side 1B and FIG. 1B according to the acquisition program for acquiring the mixing ratio and the amount of user viewing information stored in the storage device 312. As described above, based on the mixing ratio change program that changes the pointer position of the slide bar and the acquired mixing ratio information, the user-side preference ranking and the server-side preference ranking are mixed to generate new ranking information. A program and an output program for outputting the generated new ranking information. .

  Each process is realized by the processor 301 executing these programs 306 stored in the memory 305. However, these can also be realized by hardware by integrating them as a processing unit for performing each processing.

  The server-side program 318 includes a reception program that receives viewing information transmitted from a plurality of user terminals, a management program that collectively manages the viewing information, and generates server-side preference ranking information from the viewing information. And a transmission program for transmitting the overall preference ranking information to the user terminal.

  In addition, when it is set as the structure which produces | generates new ranking information on the server side, the receiving program which receives separate preference ranking information from a user terminal, the receiving program which receives mixing ratio information from a user terminal, and user side preference ranking And a server-side preference ranking and a new ranking generation program for generating new ranking information.

  Each process is realized by the processor 313 executing these programs 318 stored in the memory 317. However, these can also be realized by hardware by integrating them as a processing unit for performing each processing.

FIG. 4 is an example of a hardware configuration diagram of the TV 202 of FIG.
Reference numeral 401 denotes an antenna for receiving a broadcast signal, 402 denotes a broadcast receiving device, 403 denotes a display device, 404 denotes a control device, 405 denotes an input device such as a remote controller, and 406 denotes the Internet.

  The broadcast signal received by the antenna 401 is sent to the broadcast receiving device 402, and the user's favorite channel can be received by the display device 403 using the input device 405. Further, digital content can be received from the Internet 406 by a method such as streaming and downloading and displayed on the display device 403. The control device 404 is operated by the input device 405, and the user-side preference and the server-side preference ratio are selected from the input device 405.

  When the user instructs the user preference ranking result display from the input device 405, the control device 404 transmits a mixed ratio of user preference and server-side preference and an analysis request through the Internet 406. The result sent from the server side is displayed on the display device 403. The input device 405 may use a mouse, a keyboard, a joystick, a game controller, etc. in addition to the remote controller.

FIG. 5 is an example of a sequence diagram of a preference analysis procedure on the user side (terminal side) and server side (unspecified majority side).
Reference numeral 501 denotes a client side sequence. Reference numeral 502 denotes a sequence on the server side. Hereinafter, the sequence will be described with reference to numbers.

  In 503, the client 501 accumulates the viewing history / search history / operation history. The client 501 sends the viewing history / search history / operation history data to the server 502 at 504, and the data is accumulated at 505.

  Subsequently, the client 501 instructs the server 602 to request a ranking information analysis result at 506, and then transmits the ranking data mixture ratio to the server 502 at 607. The server 502 performs preference analysis in 508 from the information of the viewing history / search history / operation history data 506 and the mixing ratio 507 and the accumulated information 505, and sends the ranking information analysis result to the client 501 as 509. The client 501 displays the received analysis result 509 at 510 at the result.

  Thereby, the client can know not only user personal preference information but also preference information of an unspecified number of users.

FIG. 6 is an example of a viewing data accumulation table of users and servers.
601 is a viewing / reservation data number managed by the user side device, 602 is a viewing program name and reserved program name on the user side, 603 is a program start time, 604 is viewing (S) on the user side, reservation ( Y) is shown. 605 is the number of the viewing / reservation data managed on the server side, 606 is the name of the viewing program and reserved program of the entire user, 607 is the program start time, 608 is the program code such as G code, 609 is the entire It is a ratio indicating what percentage of the users watched / reserved.

  In this example, numbers are assigned in order from the oldest viewing data, but the reverse is also possible. This data represents a history of what programs the user has watched and reserved at what time.

For example, in the case of 1 in the viewing / reservation data number 601, it can be seen that the user has watched the program “Sport A” from 22:30 on October 9, 2007.
As a result, it is possible to know when and what kind of program was viewed / reserved. Each of these programs is sometimes called content.

FIG. 7 is another example of the user viewing and search program data storage table.
Reference numeral 701 denotes a viewing program data number, 702 a viewing program name, 703 a viewing time (seconds), 704 a viewing count, and 705 a search count.

  For example, in the case of 1 in the viewing reservation data number 701, it can be seen that the user has watched a program called sports A for a total of 3,500 seconds, the number of times of viewing is 30, and the number of searches is 6. Based on this information, it is converted into points for creating a ranking. As a point calculation method, for example, a product of viewing time, number of times of viewing, and number of times of search can be considered.

  As a result, the viewing time, the number of viewing times, and the number of searches for the viewing program can be known, and it can be used as a material for determining which program is being watched more.

  FIG. 8A is an example of a table in which viewing history and search history are pointed. FIGS. 8B and 8C are examples of a ranking table (b) and a preference type (c) for each genre created from this table.

  Reference numeral 801 represents a viewing history and search history ranking, 802 represents a viewing history and search history ranking, and 803 represents a genre of search keyword ranking. Reference numeral 804 denotes the number of points for ranking. 805 is a program genre, 806 is a radar chart indicating a user's preference type, 807 is a ranking of genre ranking, 808 is a genre name, and 809 is a point number for determining a ranking ranking by genre. .

  For example, in the case of # 1 in the number 801 of the viewing reservation data, the user often views and searches for keywords related to the genre of sports, and the number of points is the most as 099000930 points. You will find that you are very interested in.

  Points 804 are calculated based on the viewing time, the number of times of viewing, the number of times of search, etc. in FIG. For example, the point 804 is obtained from the total viewing time 703 × viewing count 704 × search count 705. Note that only one value of each of the total viewing time, the number of viewings, and the number of searches may be set as the point 804, or the point 804 is calculated from these combinations, for example, the total viewing time 703 × the number of viewings 704 is set to the point 804. Also good.

  As a result, it is possible to quantitatively determine which genre of the program is often watched. Further, in the ranking by category (c), the ranking is determined by totaling the number of points of programs of the same category, but other calculation methods may be used. Based on the number of points, a radar chart of preference type (b) is created. Specifically, a radar chart is created with each genre 803 as an axis and a point 804 as a value.

  In the example of FIG. 8, the user's viewing history and search history are pointed to obtain the preference type. On the other hand, the server 204 collects the viewing data in FIGS. 6 and 7 or the viewing data after the points in the table (a) in FIG. 8 from other PCs 201 and other TVs 202. In addition, when a TV program is distributed from a server by IPTV or VOD (Video On Demand), viewing data such as 605 to 609 in FIG. 6 is collected on the server.

  By obtaining the points for each genre from the viewing time, the number of times of viewing, the number of searches, and the like of these viewing data, and summing the number of points of each user for each genre, preference information for a plurality of viewers is obtained. Thereby, ranking information for the entire user is created and stored in the auxiliary storage device 319. The stored viewing history / search history is a table similar to that shown in FIGS.

  The user preference is “individual preference”, the user preference ranking is “individual preference ranking”, the server preference that is the preference of multiple users gathered in the server is “overall preference”, and the overall ranking obtained from the server preference Is sometimes referred to as “overall preference ranking”.

FIG. 9 shows an example of a method of feeding back the result inferred by the inference system from the user's viewing history.
901 is a user's viewing history and operation history, 902 is a preference extracting unit that extracts the user's preference from the user's viewing history and operation history, and 903 is a user's preference based on a result selected by the user. This is an algorithm for feeding back whether or not a program along the line has been recommended, and 904 represents the result. 905 is a result of whether or not the user has selected a program recommended by the system, and 906 is a part that feeds back the selection result to server program recommendation.

  The viewing / operating history 901 is applied to the preference extraction unit 902, ranking information is created by the adjustment algorithm 903, and the result 904 is displayed. It is determined in 905 whether the user has selected the recommended result, and the result is sent to the program recommendation system 906 of the server. The result is again applied to the adjustment algorithm 903 to perform feedback. In this way, it is possible to recommend content closer to the user's preference by reflecting whether or not the user has selected the recommended result.

  FIG. 11 is an example of a flowchart for the user terminal to determine the priority ratio between the user preference ranking and the server preference ranking. A slide bar in which the user adjusts the user preference (terminal preference) and the server preference is shown in FIG.

  When the process is started at 1101, the user terminal first accumulates user viewing information at 1102 and confirms whether or not the user viewing information is sufficiently accumulated at 1103. In 1104, it is determined whether the user's viewing information is sufficient. If the user's viewing information is sufficient, a value set by the user using the slide bar is read in 1105. Whether or not the user viewing information is sufficiently accumulated is determined according to the amount of operation history information of the user and the accumulated ratio as described above.

  Subsequently, the user terminal determines a mixing ratio of the user ranking information in 1106 from the value of the slide bar read in 1105. If the slide bar in FIG. 1 is at the right end, the server ranking is 100%, and if the slide bar is in the middle, the mixing ratio of the user preference ranking and the server preference ranking is 50%.

  If there is an analysis result request in 1107, the server performs preference analysis 1108, creates a preference ranking of the server, and transmits it to the user terminal. The server preference ranking may be created in advance. The user terminal obtains the server preference ranking at 1109, mixes the user preference and the server preference at 1110 at the mixing ratio determined at 1106, and if the rankings overlap, the user preference ranking and the server preference ranking Are compared, and a duplicate program is detected in 1111. Only one duplicate program is left and the other is deleted.

  The user terminal recommends the content ranking in 1112 and stores the recommendation result in 1113. When the user selects a recommended program, information on the selected program is transmitted to the server. In 1115, the process ends.

  Thereby, it becomes possible to provide the user with a mixed ranking of the user-side preference information and the server-side preference information at the mixing ratio designated by the user.

  Although the above process is described as being performed by the user terminal, the process may be performed by either the server or the user terminal. In this flow, the description was made without distinguishing between the content executed on the server and the content executed on the user terminal. However, in the case of corresponding to the block diagram of FIG. 9, the processing is divided as indicated by the broken line, the processing performed by the server is 906, and the other processing 901 to 905 is processing at the user terminal.

  When the server performs the above process, the server receives the mixing ratio in 903, and the server's preference ranking held by the server and the user's preference ranking received from the user terminal are mixed by the server. Then, the newly generated ranking information is notified to the user terminal.

FIG. 10 is an example of a method of determining a tendency recommended by the user terminal.
1001 is a slide for adjusting whether the recommendation is made according to the user's preference or the recommendation different from the user's preference, and 1002 is a program for recommending a program according to the user's preference. When a program with a different preference is recommended, 1004 is a setting bar.

  When the user wants to make a recommendation according to his / her preference, the slide 1001 is moved to the left, and when the user wants to make a recommendation different from his / her preference, the slide 1001 is moved to the right. Thereby, it becomes possible to specify on the user side whether or not to make a recommendation in accordance with the user's preference.

  A method for mixing preference ranking data on the user side and the server side when the user has set that the user wants to watch a program other than his / her preference will be described with reference to FIG.

  1201 is ranking data on the user side, 1202 is ranking data on the server side, that is, unspecified majority user side, and 1203 is ranking data after mixing user-side ranking data and unspecified number of users ranking data. 1204 and 1205 are mixing ratios.

  Reference numeral 1206 denotes a program in which the ranking data on the user side and the ranking data on the server side match, and the program with an underline added matches. A program 1207 is ranked in the ranking data on the user side.

  The ranking data 1201 is created based on the user's viewing history and reservation history. The ranking data 1202 is created from the result of integrating all viewing histories and reservation histories of unspecified users held by the server. The ranking data 1203 is created based on the mixing ratio 1204 and the mixing ratio 1205 by comparing the character strings of the ranking data 1201 and the ranking data 1202 to avoid duplication of programs.

  In the case of the mixing ratio in this example, since the mixing ratio 1204 is 10% and the mixing ratio 1205 is 90%, the user is not a program close to his / her preference, and he / she does not know and other unspecified many users. It seems that they want to know programs that are close to their tastes.

  Specifically, nine programs from 01 to 09, which are 90% of the ranking data on the server side, enter the ranking data 1203, and when considering the remaining 10%, 10% from the ranking data 1201 on the user side When extracting, since soccer 01 of 01 does not overlap, it is registered in the ranking as 1207.

  However, this is an example, and the remaining 10% can be extracted randomly from the ranking data on the user side, or the remaining 10% can be extracted from the lower rank of the ranking data. Ranking information may be generated by weighting the ranking information points of the server and the user terminal and mixing them. For example, in the ranking (c) by genre in FIG. 8, the user likes sports, variety, and culture in this order, but lowers these weighting factors, and conversely, drama, music, news, etc. that are not so much preferred. Increase the weighting coefficient.

  Then, by applying this weighting coefficient to each genre of the server ranking 1202, a program that is distant from the user's preference can be presented. By changing the weighting coefficient according to the ratio of the slide bar in FIG. 10, it is possible to adjust whether the recommendation according to the user's preference is performed or the recommendation different from the user's preference is performed.

  As a result, it is possible for the user to adjust whether the user wants to recommend a program close to his / her preference or whether he / she wants to recommend a program close to the preferences of an unspecified number of users.

  FIG. 13 is a display example of ranking data on the user & server side.

  Reference numeral 1301 denotes a TV viewing screen including IPTV, VOD (Video On Demand), and cable television. Reference numeral 1302 denotes a setting button for changing the ranking information display method and screen settings. Reference numeral 1303 denotes a ranking of programs recommended in FIG. 12 (after mixing of user and server). Reference numerals 1304 and 1305 denote CM (Commercial Message) and PV (Promotion Video) selected from genres with high points of viewing history / search history ranking in FIG. Reference numeral 1306 denotes a program currently being viewed. Reference numeral 1307 is a server side, that is, ranking of an unspecified number of users, and 1308 is a ranking of users.

  The user checks the ranking 1303 while viewing the program 1306 currently being viewed. The ranking 1307 of unspecified number of users and the ranking 1308 of users can be understood at a glance. The ranking 1303 may be not only a program name but also a still image or a moving image. If there is content that the user is interested in, it can be viewed and reserved from there.

  If you are concerned about CM / PV, choose CM1304 or PV1305. Then, actions such as displaying the details of the CM / PV and jumping to the Web page occur. When the setting button 1302 is selected, the user can decide which screen is displayed / hidden and at which position the screen is displayed.

  The type of content to be viewed may be not only TV but also VOD (Video On Demand) type such as download / streaming. Further, the image is not limited to a moving image and may be a still image.

  Accordingly, the user can check the ranking while viewing the program currently being viewed, and can also view the CM / PV. In addition, since the CM and PV of the genre that the user is interested in can be viewed, it is possible to provide information more effectively than simply streaming CM / PV.

FIG. 14 is another example of a display example of ranking data on the user side and the server side.
1401 is a recommended program (CM / PV), 1402 is a recommended program ranking obtained from the server-side preference result in FIG. 12, 1403 is a recommended program ranking obtained from the user-side preference result in FIG. Reference numeral 1404 denotes a setting button.

  While viewing the recommended program (CM / PV) 1401, the user can check the recommended program ranking 1402 on the user side and the recommended program ranking 1403 on the server side. When there is a program that interests you and you select that program, you can view or record the program. If it is currently being broadcast, the character color of the program name may be changed, or it may be enclosed in a square. The ranking information may be displayed not only in characters but also as a still image or a moving image.

  The recommended program (CM / PV) 1401 may be a still image or a moving image. However, in the case of a still image, the program is switched to another still image at a certain time. In the case of a video, when one video content ends, the video is switched to another video. This recommended program is selected and broadcasted based on user preference information and unspecified number of user preference information. The adjustment of the display method, the genre of the program to be recommended, and the user preference mixing ratio on the user side and the server side can be adjusted not only automatically but also from the setting side 1404 from the user side.

  If the slide bar is set to 100% server preference in FIG. 1, the recommended program ranking is only the recommended program 1403 obtained from the server-side preference result. Conversely, if the slide bar is set to 100% user preference, Only the recommended program 1402 obtained from the user's preference result is obtained. When the slide bar is not at both ends, 1402 and 1403 are mixed.

As a result, CMs and PVs that are of interest to the user can be provided to the user, and information can be provided more efficiently than simply showing the user the CM and PV.
FIG. 15 shows an example when the setting button shown in FIGS. 13 and 14 is pressed.
1501 is a radio button for selecting whether or not to recommend a program with a large number of viewing / reserving times, 1502 is a spin button for determining whether to recommend a program up to the first rank in the ranking data, and 1503 is a recommendation system. It is a recommended ranking obtained as a result.

  1504 is a radio button for selecting whether or not a program with a high number of searches is recommended, and 1505 is ranking information rearranged in descending order in the viewing history / search history ranking of FIG. Reference numeral 1506 denotes a spin button for determining whether or not a program with the highest number of searches is recommended. Reference numeral 1507 denotes a button for indicating the analysis result of the user's preference type.

  Whether the user wants to recommend a program with a large number of times of viewing / reserving as a program recommendation method or a program including a keyword with a large number of searches by selecting the radio button 1501 or the radio button 1504 Can be selected. Also, it is possible to select from the spin button 1502 and the spin button 1506, respectively, how many times the program is recommended. A detailed analysis result can be confirmed by selecting a button 1507, and can be used as a judgment material for setting.

  Thereby, the information used for program recommendation can be changed. Also, the user can set the number of programs that the user wants to recommend while viewing the ranking result.

  FIG. 16 is a display example of a user preference type analysis result. When the button 1507 in FIG. 15 is pressed, the user's preference type analysis result is displayed.

  Reference numeral 1601 represents the previous user preference analysis result, and 1602 represents the current user preference analysis result. Reference numeral 1603 represents the user's current preference type, and 1604 is a detail button.

  The previous user preference analysis result 1601 and the current user preference analysis result 1602 determine the genre from the viewing history, search history, and operation history. The preference type 1603 diagnoses the user type based on the data obtained from the analysis result and recommends a program. The user can know the tendency of his / her preference by seeing these results.

  When the user presses the detail button 1604, a more detailed graph can be displayed for each genre. In this example, sports, music, variety, news, culture, and drama are divided into six genres, but there is no problem with other genres as well. The number of genres is not limited to six.

  As a result, the user can know his / her preference type from the viewing history / search history / operation history, can grasp the current preference status, and can also visually understand the reason for recommendation from the recommendation system. . If the user does not like the recommendation result, it can be set by the user, so that it can be used as a judgment material for changing the setting.

An example of a screen displayed when the detail button 1604 in FIG. 16 is pressed will be described with reference to FIG.
1701 is a button for displaying in more detail for each genre, and is divided into sports 1701, music 1702, variety 1703, news 1704, education 1705, and drama 1706, respectively. This genre corresponds to the genre of FIG.

  In this example, sports, music, variety, news, culture, and drama are divided into six genres, but there is no problem with other genres as well. The number of genres is not limited to six. Reference numeral 1707 denotes a point display graph of keywords related to sports. These can be displayed by summing up the table (a) in which the viewing history / search history in FIG.

  When the user presses the sports button 1701, a more detailed ranking of keywords related to sports is displayed. When the user presses the music button 1702, a more detailed keyword ranking for music is displayed. Similarly, when a variety button 1703, a news button 1704, a culture button 1805, and a drama button 1706 are pressed, detailed keyword rankings for each genre are displayed.

In the graph 1707, the higher the point, the longer the length of the graph. In this example, it can be seen that the user frequently views, reserves, and searches for a program related to soccer in the genre of sports.
The user can know more detailed ranking data for each genre by selecting the button for each genre.

A screen example when the preference type analysis result button 1507 of FIG. 15 is pressed will be described with reference to FIG.
1801 is a graph showing the user's current preference analysis result, 1802 is a graph indicating which field the current recommendation system is recommending to the user, 1903 is the current user's preference type, 1804 is a slide indicating whether a user wants to recommend a program that he / she does not know, and 1805 recommends a program whose genre is different from his / her preference, whether the user wants to recommend a program whose genre is similar to his / her preference. It is a slide that you want. Reference numeral 1806 denotes a button for displaying detailed data for each genre shown in FIG.

  Reference numeral 1801 indicates a point chart (a) of the user's viewing history / search history displayed on the radar chart, similarly to your preference type (b) in FIG. On the other hand, 1802 displays the recommended ranking information on the radar chart, and by displaying these on the same screen, it is possible to confirm the current user viewing status and the recommended status in the recommendation system at a time. There are benefits.

  First, the user confirms his / her current preference analysis result 1801 and confirms the graph 1802. If the user wants to recommend a genre that is completely different from his / her preference, or wants to pioneer a new program, the current setting, slide 1804, is closer to the new program pioneer, and slide 1805 is You just have to be close to the genre different from your preferences.

  On the other hand, if you want to recommend a program in a genre that is similar to your taste, or if you want to thoroughly pursue the program or genre you are watching, make slide 1804 closer to the pursuit of preference. The slide 1805 may be changed to a genre similar to the preference. By changing the settings of the slide 1804 and the slide 1805, the priority items of the recommendation system change, and the recommendation similar to the user's preference type is performed.

  Since the recommendation status 1802 of the recommendation system changes according to the movement of these slide bars, it is possible for the user to determine what kind of recommendation status is achieved by moving the slide bar.

  Thereby, according to the user's preference, the user can freely set whether to recommend a program of a genre close to the user's preference or to recommend a completely different program. Since this recommendation system can be implemented without being connected to a network, it can be implemented simply by incorporating the system into the terminal (within the terminal). When connecting to a network, it is possible to perform more flexible program recommendation by adding unspecified number of user preference ranking information other than users to programs recommended by the recommendation system.

  A procedure for changing the genre of a program recommended by the recommendation system when the user moves the slide bar in FIG. 18 will be described with reference to FIG. Processing is started in 1901 and an application is started in 1902. In 1903, it is determined whether or not the user has moved the slide bar. If the user has moved, the position information of the slide bar at that time is acquired in 1904.

  In 1905, user preference information (number of points) is acquired. In 1906, the ratio of the user-side preference and the unspecified number of preferences is calculated based on the position information of the slide bar. In 1907, the calculation result is stored. Based on this result, the preference type is determined in the same manner as in FIG. 11 in 1908, a radar chart is created in 1909, and the result is output on the screen in 1910. In 1911, the process ends.

An example of a screen when the preference type analysis result button 1507 in FIG. 15 is pressed will be described with reference to FIG.
2001 is a user's preference type, 2002 is a preference type for all unspecified users, and 2003 represents a bias of a genre recommended by the recommendation system. 2004 is a slide indicating whether a user wants to recommend a program that he / she does not know, and 2005 indicates that a user wants to recommend a program of a genre similar to his / her preference, or recommends a program of a genre which is completely different from his / her preference. It is a slide of what you want. 2006 represents a slide in the time zone of the program that you want to recommend. Reference numeral 2007 denotes a button for displaying detailed data for each genre, similar to 1806 in FIG.

  First, the user confirms his / her current preference state 2001, confirms the preference type 2002 of the entire unspecified number of users, and confirms the current recommendation system status 2003. If the user wants to recommend a genre that is completely different from his / her preference, wants to pioneer a new program, or wants to know a program to be aired at noon, the current setting is 2004. It is only necessary to move toward the new genre pioneering group, make the slide 2005 close to the genre, and move the slide 2006 around the daytime program.

  On the contrary, “I want you to recommend a program of a genre similar to my taste”, “I want to thoroughly pursue the program / genre I am watching”, “I want to know the program that will be broadcast at night” For example, the slide 2004 may be made close to a preference pursuit group, the slide 2005 may be changed to a genre similar to the preference, and the slide 2006 may be made closer to a night program. By changing the settings of the slide 2004, the slide 2005, and the slide 2006, the priority items of the recommendation system change, and the recommendation according to the user's preference type is performed.

  Thereby, according to the user's preference, the user can freely set whether to recommend a program of a genre close to the user's preference or to recommend a completely different program.

  FIG. 21 is a flowchart for implementing FIG. The processing is started at 2101, the points of FIG. 8 are added at 2102, and the ratio of each genre to all viewing programs is calculated at 2103, for example. In 2104, user preference information (calculation result) is stored, and in 2105, the setting of the recommendation system is confirmed.

  Here, the position of the slide is acquired, the recommendation ratio of the recommendation system is calculated in 2106, and the result is stored in 2107. Here, whether to recommend a user preference, whether to recommend a new genre program, or what time to recommend a program is determined from the values of the slide bars 2004-2006. In 2108, it is confirmed whether or not connected to the network. If connected, user preference information is transmitted to the server in 2109, and preference information of an unspecified number of users is acquired in 2110.

  It is determined whether information can be acquired in 2111. If it can be acquired, preference information of an unspecified number of users is stored in 2112. Based on the user preference information obtained in 2104, information on the recommendation system obtained in 2107, information on an unspecified number of users obtained in 2112, a radar chart is created in 2113, displayed on the screen in 2114, Program recommendation is performed, and the process ends at 2115.

  If it is not connected to the network or if the preference information of an unspecified number of users cannot be acquired, a radar chart is created from only the user preference information in 2113, displayed on the screen in 2114, and program recommendation is performed. In 2115, the process ends.

However, the program to be considered does not have to be an all-viewing program, for example, 2103 is a viewing program for the past year, or when it exceeds a certain capacity, the old program is not considered.
FIG. 22 shows a screen example when the preference type analysis result button 1507 of FIG. 15 is pressed.
2201 represents the previous preference type of the user, 2202 represents the current preference type of the user, 2203 represents the previous preference type of the server side, that is, the unspecified number of users, and 2204 represents the current preference type of the unspecified number of users. Represents a type. Reference numeral 2205 denotes a button for displaying detailed data for each genre in FIG.

  The user shows the result 2201 of the user when viewing the previous preference type analysis result, the result 2203 of the unspecified majority user, the result 2202 of the user when viewing the current preference type analysis result, and the results of the unspecified majority user. By comparing the results of the result 2204, it is possible to confirm how the user's preference and the preferences of the unspecified number of users have changed. Thereby, the preference type of each genre can be objectively known.

  FIG. 23 shows a screen example when the preference type analysis result button 1507 of FIG. 15 is pressed.

  2301 represents the previous preference type of the user, 2302 represents the current preference type of the user, 2303 represents the previous preference type of the server side, that is, the unspecified number of users, and 2304 represents the current preference type of the unspecified number of users. Represents a type. 2305 and 2306 are tabs for displaying the preference type for each month, 2305 is a tab for January, and 2306 is a tab for December. Even in a month in which the user does not see the preference type result, for example, when the month changes to a new month, the preference type result of the user side and the unspecified majority user side may be automatically saved.

When the user sees the result 2301 of the user when viewing the previous preference type analysis result, the result 2303 of the unspecified number of users, and the current preference type analysis result in this month (January in this example) The user result 2302 can be compared with the result of the unspecified majority user result 2304 to confirm how the user's preference and the preference of the unspecified majority user have changed. In order to check the preference type analysis result of last month, the December tab 2306 may be selected.
Thereby, a user's preference type transition can be retroactively confirmed on a monthly basis.
FIG. 24 shows an example of a setting screen for recommending not only the program recommendation of FIG. 15 but also a Web page.
2401 is a radio button for determining whether or not to recommend a Web page with a large number of visits by a user, 2402 is a spin button for determining whether or not a Web page with the highest number of visits is recommended, and 2403 The recommendation ranking obtained as a result of the recommendation system.

  2404 is a radio button for selecting whether to recommend a Web page with a high number of searches, 2405 is a spin button for selecting whether to recommend a Web page with a high number of searches, and 2406 is the viewing of FIG. The ranking information is sorted in descending order of points in the history / search history ranking. Reference numeral 2407 denotes a button for indicating the analysis result of the user's preference type.

  The user selects either the radio button 2401 or the radio button 2404, and as a web page recommendation method, the user wants to recommend a web page with a large number of viewing / reserving times, or the radio button 2404 has a large number of searches. It is possible to select whether a Web page including a keyword is recommended.

  In addition, it is possible to select from the spin button 2402 and the spin button 2405, respectively, how many Web pages the number of times is recommended. As information for selection, recommended ranking 2403 and ranking information 2406 are displayed, and the user selects which to select based on the ranking. A more detailed analysis result can be confirmed by selecting a button 2407, and can be used as a judgment material for setting.

  The information used for selection can be changed depending on which of the radio button 2401 and the radio button 2404 is selected. From the recommended ranking 2403 and the ranking information 2406, it is possible to set the number of Web pages to be recommended from the user side while viewing the ranking result.

By taking the configuration of this embodiment, the following effects are obtained.
1) Even when the user's preference information cannot be acquired in the initial state (when there are few keywords), it can be supplemented and recommended by an unspecified number of preference information.
2) By recommending popular content that is not recommended from user preference information, it is possible to recommend content without collecting information about the user and to prevent the user's preference from being biased. it can.
3) By recommending popular content that is not recommended based on the user's preference information, the content that matches the user's changed preference is popular with an unspecified number of people even immediately after the user's preference changes. It can be found from the content. Here, the user preference information represents the genre and keyword of the content that the user often views and reserves, and the keyword that the user often searches, and the user preference, that is, the content that the user likes is analyzed / recommended from such information. To do.
4) If the user likes an unspecified number of recommended contents, the user's preference range further expands, so that the range of his / her favorite genre can be expanded.
5) It is possible to change the ratio on the user side, whether the preference on the user side is recommended or the preference is given to an unspecified number of preferences.

Examples of methods for adjusting the amount of user preference and server preference information Example network configuration Example of hardware configuration diagram of user's PC (client) 205 and server 204 in FIG. Example of hardware configuration diagram of TV 202 Example sequence diagram of preference analysis procedure on user side (terminal side) and server side (unspecified majority side) Example of viewing data storage table of user and server Another example of user viewing data accumulation table Example of table (a) with viewing history and search history as points, your preference type (b), ranking by genre (c) Example of how to feed back the results inferred by the inference system from user viewing history Example of how to determine a recommendation tendency Example of flowchart for determining priority ratio of user preference ranking and server preference ranking An example of a method of mixing user-side and server-side preference ranking data when the user has set the user to view a program other than his / her preference Display example of ranking data on the user & server side Another example of ranking data on the user side and server side Examples of the setting buttons in FIGS. 13 and 14 Display example of user preference type analysis results Example screen displayed when the detailed button 1604 in FIG. 16 is pressed. Screen example when the preference type analysis result button 1507 in FIG. 15 is pressed Example of the procedure to change the genre of the program recommended by the recommendation system when the user moves the slide bar Screen example when the preference type analysis result button 1507 in FIG. 15 is pressed Example of flowchart for implementing FIG. Screen example when the preference type analysis result button 1507 in FIG. 15 is pressed Screen example when the preference type analysis result button 1507 in FIG. 15 is pressed Setting screen for recommending web pages

Explanation of symbols

101 Slide position when user preference and server preference are approximately half 102 Slide position when user's preference is sufficiently collected 103 Setting bar 201 PC (client) of other user 1
202 TV of other user 2
203 Internet 204 Server 205 User PC (client)
206 User 1 preference analysis result 207 User 2 preference analysis result 208 PC (client) and TV preference analysis result 209 User preference analysis result 210 208 Combined preference analysis result of 208 and 209 preference analysis result ( Logical AND)
301 CPU
302 TV (content) viewing application 303 Driver 304 Interface hardware 305 Memory 306 Program 307 Display device 309 Mouse 310 Keyboard 308 Input device 311 Internet 312 Auxiliary storage device 313 Server CPU
314 Server application 315 Server driver 316 Server interface hardware 317 Server memory 318 Server program 319 Server auxiliary storage device 401 Antenna for receiving broadcast signal 402 Broadcast receiving device 403 Display device 404 Control device 405 Remote control etc. Input device 406 Internet 501 Client side sequence 502 Server side sequence 503 History accumulation 504 History data transmission 505 Data accumulation 506 Ranking information analysis result request 507 Mixing ratio data transmission 508 Preference analysis 509 Analysis result 510 Result display 601 User side device No. of viewing / reservation data managed in 602. Name of viewing program on user side and reserved program name 603 Program start time 604 Viewing on user side (S), reservation (Y) 605 Number of viewing / reservation data managed on the server side 606 Viewing name and reserved program name of the entire user 607 Program start time 608 Program code 609 such as G code What percentage of all users of viewing / reserving A ratio 701 indicating the number of viewing program data 702 Viewing program name 703 Viewing time (seconds)
704 View count 705 Search count 801 View history and search history ranking 802 View history and search history ranking 803 Search keyword ranking genre 804 Ranking points 901 Viewing history and operation history 902 Inference system 903 Adjustment algorithm 904 Result 905 Recommended program selection result 906 Server program recommendation 1001 Slide 1002 Program recommendation 1003 according to user's preference 1003 Program recommendation different from user's preference 1004 Setting bar 1101 Start of processing 1102 User preference accumulation 1103 User operation information collection amount confirmation 1104 User operation information Amount determination 1105 Set with slide bar 1106 Mix ratio determination 1107 Server preference analysis result request 1108 Preference analysis 1109 Server preference ranking acquisition 1110 User preference and server Preference mixing 1111 Duplicate ranking detection 1112 Ranking recommendation 1113 Recommendation result storage 1201 User-side ranking data 1202 Server-side (unspecified many-user side) ranking data 1203 After user-side ranking data and unspecified many-user ranking data are mixed Ranking data 1204 Mixing ratio 1205 Mixing ratio 1206 Program 1207 in which the ranking data on the user side and the ranking data on the server side match The program 1301 that is ranked in the ranking data on the user side 1301 TV viewing screen 1302 Setting button 1303 Recommended program Ranking (after mixing user and server)
1304 Recommended CM (Commercial Message)
1305 Recommended PV (Promotion Video)
1306 Program currently being viewed 1307 Ranking on server side (unspecified number of users) 1308 Ranking on user side 1401 Recommended programs (CM / PV)
1402 User-side recommended program ranking 1403 Server-side recommended program ranking 1404 Setting button 1501 Radio button 1502 for selecting whether or not to recommend a program with a large number of viewing / reserving times. Spin button to determine 1503 Recommended ranking obtained as a result of the recommendation system 1504 Radio button for selecting whether to recommend a program with a large number of searches 1505 Point order ranking information 1506 How many programs with the highest number of searches are recommended Spin button 1507 for determining whether or not User preference type analysis result display button 1601 Previous user preference analysis result 1602 Current user preference analysis result 1603 Current user preference type 1604 Detail button 1701 Detail button “Sport”
1702 Detail button "Music"
1703 Detail button "Variety"
1704 Detail button "News"
1705 Detail button "Liberal Arts"
1706 Detail button "Drama"
1707 Point display graph of keywords related to sports 1801 User's current preference analysis result 1802 Recommendation status display graph by recommendation system field 1803 Current user preference type 1804 Slide 1805 Slide 1806 Detailed data display button by genre 1901 Start processing 1902 Start application 1903 Slide bar movement determination 1904 Slide bar position information acquisition 1905 User preference information acquisition 1906 Recommendation system recommendation ratio calculation 1907 Storage of recommended system information (calculation result) 1908 Preference type determination 1909 Radar chart creation 1910 Display on screen 1911 Processing End 2001 Preference type 2002 for all unspecified number of users User preference type 2003 Slide 2004 Slide 2005 Slide 200 6 Detailed data display button for each genre 2101 Start of processing 2102 Add points for each genre 2103 Calculate the ratio of each genre to the viewing program 2104 Save user preference information (calculation result) 2105 Check the recommended system settings 2106 Calculate the recommended ratio of the recommended system 2107 Storage of recommended system information (calculation result) 2108 Determination of online connection 2109 Transmission of user preference information to server 2110 Acquisition of preference information of entire user 2111 Success / failure determination of information acquisition 2112 Storage of preference information of entire user 2113 Radar chart creation 2114 Display on screen 2115 End of processing 2201 User's previous preference type 2202 User's current preference type 2203 Previous preference type 2204 on server side (unspecified large number of users) Current preference on server side (unspecified large number of users) Type 2205 Detailed data display button by genre 2301 User's previous preference type 2302 User's current preference type 2303 Server-side (unspecified majority user) previous preference type 2304 Unspecified majority user's current preference type 2305 January tab 2306 December tab 2401 Radio button 2402 indicating whether or not to recommend a Web page with a high number of visits 2402 Spin button 2403 for determining whether or not to recommend a Web page with the highest number of visits 2403 Recommended ranking 2404 Search count Radio button 2405 for selecting whether or not to recommend a Web page with a large number of spin buttons 2406 for selecting whether or not to recommend a Web page with a large number of searches Ranking information 2407 sorted in descending order of points 2407 User preference type analysis Result display button

Claims (20)

A content recommendation method by a content recommendation device,
User viewing information is stored in a storage device,
Generate individual preference ranking information indicating user preferences from the user viewing information,
Obtaining overall preference ranking information indicating the preferences of the plurality of users obtained from the viewing information of the plurality of users,
Obtaining mixing ratio information indicating a ratio of mixing the user preferences and the plurality of user preferences;
Based on the mixing ratio information, the individual preference ranking and the overall preference ranking are mixed to generate new ranking information,
A content recommendation method for outputting the new ranking information. The content recommendation method according to claim 1,
The viewing information includes at least one of viewing time information, viewing count information, and search count information for each content, and is based on at least one of the viewing time information, the viewing count information, and the search count information. A content recommendation method characterized by generating an individual preference ranking. The content recommendation method according to claim 1,
The viewing information includes two or more of viewing time information, viewing count information, and search count information for each content, and the individual preference ranking is generated based on point information calculated based on the information. A content recommendation method characterized by: The content recommendation method according to claim 1,
A content recommendation method, characterized in that a radar chart is output with the point information calculated based on viewing information of each content as a value, with the genre of each content as an axis. The content recommendation method according to claim 1,
A content recommendation method, wherein a ratio of mixing the user preference and the overall preference is changed according to the amount of viewing information of the user stored in the storage device. The content recommendation method according to claim 5,
A content recommendation method, characterized in that, as the amount of viewing information of the user stored in the storage device increases, a mixture ratio of overall preferences is decreased. The content recommendation method according to claim 1,
A content in which the ratio of mixing the user preference and the overall preference is changed according to the amount of the user viewing information stored in the storage device with respect to the total amount of viewing information that can be stored in the storage device. Recommendation method. The content recommendation method according to claim 1,
A content recommendation method, wherein information for changing a ratio of mixing the user preferences and the plurality of user preferences is displayed on a screen. The content recommendation method according to claim 1,
Content that overlaps between the overall preference ranking and the individual preference ranking is removed from the overall preference ranking to generate new ranking information,
A content recommendation method for outputting the new ranking information. The content recommendation method according to claim 1,
The content recommendation method, wherein the individual preference ranking and the new ranking information are output on the same screen. A content recommendation device,
A storage device for storing user viewing information;
An individual preference ranking generation unit that generates individual preference ranking information indicating user preferences from the user viewing information;
A receiving unit for receiving overall preference ranking information indicating the preferences of the plurality of users obtained from viewing information of the plurality of users;
An acquisition unit that acquires mixing ratio information indicating a ratio of mixing the user preferences and the plurality of user preferences;
Based on the mixing ratio information, a new ranking generation unit that mixes the individual preference ranking and the overall preference ranking to generate new ranking information;
An output unit for outputting the new ranking information;
A content recommendation device comprising: The content recommendation device according to claim 11,
The viewing information includes at least one of viewing time information, viewing count information, and search count information for each content, and the individual preference ranking generation unit includes the viewing time information, the viewing count information, and the search count information. A content recommendation device that generates the individual preference ranking based on at least one of the following. The content recommendation device according to claim 11,
The viewing information includes two or more of viewing time information, viewing count information, and search count information for each content, and the individual preference ranking generation unit is based on point information calculated based on these pieces of information. A content recommendation device that generates the individual preference ranking. The content recommendation device according to claim 11,
The content recommendation device, wherein the output unit outputs a radar chart having the genre of each content as an axis and having point information calculated based on viewing information of each content as a value. The content recommendation device according to claim 11,
A mixing ratio changing unit that changes a ratio of mixing the user preference and the overall preference according to the amount of the viewing information of the user stored in the storage device;
A content recommendation device comprising: The content recommendation device according to claim 15,
The content recommendation apparatus, wherein the mixing ratio changing unit decreases the mixing ratio of overall preferences as the amount of viewing information of the user stored in the storage device increases. The content recommendation device according to claim 11,
A mixing ratio changing unit that changes a ratio of mixing the user preference and the overall preference according to the amount of the viewing information of the user stored in the storage device with respect to the total amount of viewing information that can be stored in the storage device;
A content recommendation device comprising: The content recommendation device according to claim 11,
The content recommendation apparatus, wherein the output unit displays information for changing a ratio of mixing the user's preference and the plurality of user's preferences on a screen. The content recommendation device according to claim 11,
The new ranking generation unit generates new ranking information by excluding content overlapping between the overall preference ranking and the individual preference ranking from the overall preference ranking,
The content recommendation device, wherein the output unit outputs the new ranking information. The content recommendation device according to claim 11,
The content recommendation device, wherein the output unit outputs the individual preference ranking and the new ranking information on the same screen.

Images