JP4952348B2 - Content recommendation device and content recommendation program - Google Patents

Description

  The present invention relates to a content recommendation device and a content recommendation program for recommending content desired by a user.

  In recent years, with the development of network infrastructure, users can select and view desired content from a variety of content via a network when trying to acquire content such as music. However, on the other hand, since the number of contents is enormous, it becomes difficult for users to comprehensively understand what kind of content is managed and where it is possible to select content that suits their preferences. It was difficult.

  Therefore, a technique for recommending content to a user using collaborative filtering, which is a method for estimating the target user's preference based on preference information of a user having a preference similar to that of a certain user, is known. It has been. For example, a collaborative filtering system is known in which user preference information is centrally managed by a server, and processing such as preference prediction and content recommendation is executed on the server.

  However, in such a collaborative filtering system, the load on the server increases, so it is necessary to prepare hardware and a network environment that can withstand a considerable load, and the cost of server operation has increased. Furthermore, when the load on the server exceeds the assumption, the server may become inoperable and content recommended for users may not be provided.

Therefore, Patent Document 1 discloses that the device 2 that transmits its own profile information from the device 1 that it uses to the device 2 that is used by another user and that receives the profile information from the device 1 Using the profile information related to the user and the received profile information, the degree of correlation between the two is calculated, and when the calculation result is transmitted to the apparatus 1, the apparatus 1 selects the communication partner apparatus having a high degree of correlation from the correlation degree calculation result. There has been proposed an information transmission system that selects and notifies or receives usage information through communication with the apparatus.
JP 2006-11550 A

  However, in the information transmission system described in Patent Document 1, a device used by a user communicates with a device used by another user to acquire the degree of correlation of profile information between them, and a device with a high degree of correlation is used. Since the screening is performed, the device used by the user needs to communicate with all the devices connected to the network. Therefore, the information transmission system described in Patent Document 1 is difficult to be applied in a large-scale network such as the Internet.

  Therefore, in the information transmission system described in Patent Document 1, attempts are made to apply it to a large-scale network by segmenting the network using a simple server and mediating communication with the simple server.

  However, in such a method using a simple server, the communication range of the device used by the user is limited due to the segmentation, so the accuracy of collaborative filtering is reduced, and content is appropriately recommended to the user. There were cases where it was impossible.

  The present invention has been made in view of the above problems, and a content recommendation device and content recommendation capable of highly accurate content recommendation for each user without using hardware or a network environment that can withstand a considerable load. The purpose is to provide a program.

In order to solve the above-described object, the first feature of the content recommendation device according to the present invention is to specify bibliographic information storage means for storing bibliographic information related to content and a content recommendation device other than the content recommendation device. Similar to the read bibliographic information for the link node information storing means for storing the link node information, and reading the stored bibliographic information and link node information, respectively, for other content recommendation devices corresponding to the read link node information Device information for specifying a content recommendation device having bibliographic information to be transmitted, and a first similarity requesting transmission of similar node information including the similarity between the bibliographic information of the specified content recommendation device and the read bibliographic information Node information request means and similar node information sent from other content recommendation devices , Based on the received similar node information is requested and the node updating means for updating the link node information stored in the link node information storing means, the transmission of the similar node information from other content recommendation device, this transmission request When the number of passed devices is equal to or less than a predetermined threshold, the link node information stored in the link node information storage unit is read, and similar to other content recommendation devices corresponding to the read link node information Second similar node information requesting means for requesting transmission of node information, similar node information requested for transmission by the second similar node information requesting means, reading bibliographic information stored in the bibliographic information storage means, If the read bibliographic information is similar to the bibliographic information included in the similar node information, device information, Adds the similarity between the read bibliographic information and the bibliographic information included in the similar node information, and generates a number of similar node information that does not exceed a predetermined upper limit as a new similar node information from a high similarity The node transmission means for transmitting the generated new similar node information to the content recommendation device that requested the transmission of the similar node information, and the link node information updated by the node update means is read from the link node information storage means, Content list requesting means for requesting another content recommendation device corresponding to the read link node information to transmit a content list that is a list of content possessed by the playback device provided with the other content recommendation device, and other content recommendation The content list sent from the device is received, and the received content list There is provided a content selection means for selecting a predetermined number of contents as recommended contents on the basis thereof, and an output means for outputting the selected recommended contents to the reproduction apparatus.

In order to solve the above object, the second feature of the content recommendation device according to the present invention is that sample content data storage means for storing sample data for trial viewing corresponding to content, and transmission of a content list are requested. The bibliographic information stored in the bibliographic information storage unit is read, a predetermined number of bibliographic information is selected from the read bibliographic information, and sample data corresponding to the selected bibliographic information is read out from the sample content data storage unit. The content response means for transmitting the read sample data and the selected bibliographic information as a content list to a content recommendation device that has requested transmission of the content list, and selecting content from the recommended content by user operation Received when a content selection signal is received for The sample data corresponding to the content indicated by the content selection signal and in that a sample output control means for outputting the content output means selected from the content list received by the content selection device.

  According to the content recommendation device and the content recommendation program according to the present invention, highly accurate content can be recommended for each user without using hardware or a network environment that can withstand a considerable load.

  The best mode for carrying out the present invention will be described below with reference to the drawings.

  A content recommendation device according to an embodiment of the present invention communicates with other content recommendation devices connected via a peer-to-peer network (hereinafter referred to as a P2P network), and recommends music as desired content by collaborative filtering. Is.

  Here, collaborative filtering is a method of estimating the target user's preference based on the preference information of another user having a preference similar to that of a certain user.

  FIG. 1 is a configuration diagram showing a configuration of a content recommendation device 1 according to the present embodiment.

  The content recommendation device 1 according to the present embodiment is connected to an audio output unit 2 including a speaker or the like, and causes the audio output unit 2 to output audio data and audio data for trial listening.

  In addition, the content recommendation device 1 according to the present embodiment includes an audio data input unit 101, a user interface unit 102, an audio data decoding unit 103, a network connection unit 104, a music capturing unit 105, and an audio data code. Conversion unit 106, music playback unit 107, music purchase processing unit 108, content recommendation unit 109, link update unit 110, search response unit 111, encoded audio data storage unit 112, and bibliographic information storage unit 113, an audio data storage unit for trial listening 114, a link node information storage unit 115, and an audio data generation unit for trial listening 116.

  The audio data input unit 101 is configured by, for example, a compact disk (hereinafter referred to as CD) drive, and has a function of inputting music data from a storage medium such as a CD in which music is stored.

  The user interface unit 102 includes an input device such as a remote controller, and generates various operation signals for content recommendation, such as a display instruction signal for music bibliographic information and a music selection signal, by a user operation. Further, the user interface unit 102 includes an image output device such as a liquid crystal display, and outputs an image based on signals received from the music capturing unit 105, the music playback unit 107, and the music purchase processing unit 108.

  The audio data decoding unit 103 converts (decodes) the encoded audio data or the audition audio data input by the music playback unit 107 or the content recommendation unit 109 into audio data, and supplies the audio data to the audio output unit 2.

  Here, the audio data is data that can output music data by the audio output unit 2 such as a speaker. The encoded audio data is data in which music data is converted into a data format suitable for storage in a storage unit or network transmission, such as MP3, for example. The data is cut out and converted to a data format suitable for high compression listening. Audio data is a general term for audio data, encoded audio data, and audio data for trial listening.

  The network connection unit 104 is a communication interface such as a LAN card. By connecting to the network 17 through the network connection unit 104, the network connection unit 104 can communicate with other content recommendation devices and servers connected to the network 17. In particular, content recommendation devices are logically connected to form a P2P network.

  When the user interface unit 102 is supplied with a signal requesting to fetch a song, the song fetching unit 105 reads the audio data of the song from the audio data input unit 101 and audio along with a song ID that uniquely identifies the song. The data is encoded by the data encoding unit 106 and stored in the encoded audio data storage unit 112. Further, the network connection unit 104 is connected to a music bibliographic information database (not shown) via the network 17 to acquire bibliographic information of the read music, and associates the acquired bibliographic information with a music ID for uniquely identifying the music. And stored in the bibliographic information storage unit 113. As the music bibliographic information database, for example, there is a database that provides bibliographic information related to a music CD called CDDB (Compact Disc DataBase).

  When the music playback unit 107 is supplied with a music ID along with a music ID from the user interface unit 102, the music playback unit 107 reads out the encoded audio data corresponding to the supplied music ID from the encoded audio data storage unit 112. The read encoded audio data is supplied to the audio data decoding unit 103. In addition, the music reproducing unit 107 adds “1” to the number of reproductions corresponding to the music ID of the read encoded audio data among the bibliographic related information stored in the bibliographic related information storage unit 113.

  When a signal instructing purchase of music is supplied from the user interface unit 102, the music purchase processing unit 108 performs music purchase processing from a content provider connected via the network 17 from the network connection unit 104. Specifically, there is a process of ordering a storage medium such as a CD in which music is stored, or a process of downloading encoded audio data of a music from a content provider. In either case, it is necessary to perform settlement processing. In addition, when downloading the encoded audio data of the music from the content provider, the bibliographic information of the music is also downloaded at the same time, a music ID that can uniquely identify the music is generated, and the generated music ID and the encoded audio data are used. The associated music ID is stored in the encoded audio data storage unit 112, and the generated music ID, the bibliographic information, and the number of reproductions are associated with each other and stored in the bibliographic related information storage unit 113 as bibliographic related information.

  The content recommendation unit 109 includes a content list request unit 109a, a content selection unit 109b, and a sample output control unit 109c in terms of its functions.

  The content list request unit 109a reads the link node information updated by the link update unit 110 and stored in the link node information storage unit 115, and sends the other content recommendation device corresponding to the read link node information to the other content recommendation device. It requests transmission of a content list that is a list of bibliographic information of music pieces and audio data for trial listening that the content recommendation device has.

  Here, the node is each content recommendation device that constitutes the P2P network, and the link node is a node in which a logical link is formed with this node. Furthermore, information for specifying a node such as an IP address related to the node is referred to as node information, and particularly, node information related to the link node is referred to as link node information. The link node information may include the degree of similarity between the bibliographic information that the node has and the bibliographic information that the link node has.

  The content selection unit 109b receives the content list requested to be transmitted by the content list request unit 109a from the link node over the network, calculates the duplication number of the music based on the received content list, and calculates the duplication number of the calculated music in descending order. A predetermined number of songs are selected as recommended contents.

  Alternatively, when the link node information stored in the link node information storage unit 115 includes the similarity between the bibliographic information of this node and the bibliographic information of the link node, the acquired content list and link node information Based on the link node information stored in the storage unit 115, the total of the bibliographic information similarity of the link node is obtained for each music piece, and a predetermined number of music pieces are selected as recommended contents in descending order of the obtained total similarity degree. To do.

  Further, the audio data for trial listening included in the content list of the selected recommended content is output to the sample output control means 109c.

  The sample output control means 109c has temporary storage means such as a memory, for example, and temporarily holds the audition audio data input from the content selection means 109b. Then, when a content selection signal for selecting music from the recommended content is received by the user's operation, the audio data for audition of the music indicated by the content selection signal is read from the temporary storage means, and the read audition data The audio data is output to the audio decoding unit 103.

  The link update unit 110 periodically operates at a predetermined time interval, and includes a first similar node information request unit 110a and a node update unit 110b in terms of its function.

  The first similar node information request unit 110a reads the link node information stored in the link node information storage unit 115 and reads the bibliographic information stored in the bibliography related information storage unit 113, and corresponds to the read link node information. To the other content recommendation device, the device information for specifying the other content recommendation device having bibliographic information similar to the read bibliographic information, and the similarity between the bibliographic information that the link node has and the bibliographic information that this node has Request transmission of similar node information.

  The node update unit 110b receives the similar node information requested to be transmitted by the first similar node information request unit 110a from the link node via the network 17 by the network connection unit 104, and based on the received similar node information. The link node information stored in the link node information storage unit 115 is updated.

  The search response unit 111 includes a second similar node information request unit 111a, a node transmission unit 111b, and a content list response unit 111c.

  As will be described later, the second similar node information request unit 111a receives supply of the transmission request (including the search query) received by the node transmission unit 111b, and receives the link node information stored in the link node information storage unit 115. Read, and request similar node information to other content recommendation devices corresponding to the read link node information.

  The node transmission unit 111b receives a transmission request (including a search query) by the first or second similar node information request unit 110a, 111a of another node.

  The node transmission unit 111b receives the similar node information requested by the second similar node information request unit 111a from the link node, reads the bibliographic information stored in the bibliography related information storage unit 113, and reads the read bibliography. If the information is similar to the bibliographic information included in the search query, the terminal information of the content recommendation device 1 and the similarity between the read bibliographic information and the bibliographic information included in the search query are added to the similar node information. The added similar node information is transmitted to the other node that has transmitted the search query transmission request signal.

  When the content list response unit 111c is requested to transmit the content list from another node via the network connection unit 104, the content list response unit 111c reads the bibliographic information stored in the bibliography related information storage unit 113, and the read bibliographic related information. The number of bibliographic information that does not exceed a predetermined threshold value is extracted from the one with the larger number of reproductions. Then, the audition audio data corresponding to the music ID of the extracted bibliographic information is read from the audition audio data storage unit 114, and the read audition audio data and the extracted bibliographic information of the bibliographic information are stored in the content list. The content list transmission request signal is supplied to the network connection unit 104 so as to be returned to the other node that has transmitted.

  The encoded audio data storage unit 112 is configured by a storage device such as a hard disk, for example, and stores a song ID that can uniquely identify a song and the encoded audio data in association with each other.

  The bibliographic information storage means 113 is configured by a storage device such as a hard disk, for example, and stores the music ID that can uniquely identify the music, the bibliographic information, and the number of reproductions as bibliographic information.

  The audition audio data storage unit 114 stores a song ID that can uniquely identify a song and trial audio data that is sample data of the song in association with each other.

  The link node storage unit 115 is configured by a storage device such as a hard disk, and stores link node information generated by the link update unit 110.

  The audition audio data generation unit 116 reads out the song ID stored in the encoded audio data storage unit 112 at a constant period, and also reads out the song ID stored in the audition audio data storage unit 114 to obtain the audition audio data. It is determined whether or not a music ID that does not exist in the music ID stored in the storage unit 114 is stored in the encoded audio data storage unit 112. If it is determined that a music ID that does not exist in the music ID stored in the trial audio data storage unit 114 is stored in the encoded audio data storage unit 112, the encoded audio data corresponding to the music ID is stored. Are read from the encoded audio data storage unit 112. Then, a part of the music of the encoded audio data that has been read out is cut out and re-encoded to generate trial audio data, and the trial audio data storage unit 114 associates the generated audio data with the music ID. Remember me.

  Note that the content recommendation device 1 according to the present embodiment includes an encoded audio data storage unit 112, a bibliographic information storage unit 113, a trial audio data storage unit 114, and a link node information storage unit 115, which are separately provided. Although configured as a storage device, each storage unit may be partially shared, or all the storage units may be shared by a single storage device.

  FIG. 2 is a configuration diagram showing a configuration of a content recommendation system to which the content recommendation device 1 according to the present embodiment is applied.

  The content recommendation system 10 shown in FIG. 2 includes a node information server 11, a plurality of content recommendation devices 12 to 16 having the same configuration as the content recommendation device 1 described above, and content providers 18 and 19 that provide content. It is connected and configured via a network 17.

  The content recommendation devices 12 to 16 are logically peer-to-peer connected to each other via the network 17 and serve as nodes forming a P2P network. In FIG. 2, the connection between the content recommendation devices indicated by the broken line indicates that the connection is a logical peer-to-peer connection. For example, the content recommendation device 12 is only connected to the content recommendation devices 13 and 15. Each indicates a peer-to-peer connection.

  In the P2P network formed in this way, the content recommendation devices 12 to 16 transmit and receive user bibliographic information and the like, thereby recommending highly accurate content by collaborative filtering.

  The node information server 11 is operated by, for example, the manufacturer of the content recommendation device 1 and stores node information regarding nodes (content recommendation devices 12 to 16) constituting the P2P network. Provides node information.

  Further, instead of operating the node information server 11 to provide node information to each node, the nodes of the selection support apparatus owned by the users are sent to general-purpose information transmission means (e-mail) on the network. , Electronic bulletin boards, etc.), and the user may input node information to the selection support apparatus as necessary.

  The content providers 18 and 19 provide a service for distributing music to the content recommendation devices 12 to 16. Specifically, the music is searched based on the bibliographic information in response to requests from the content recommendation devices 12 to 16. In addition, in response to a request from the content recommendation devices 12 to 16, a payment process for purchasing the designated music is performed, and an order for a storage medium such as a CD on which the music is stored is received or designated. The encoded audio data is distributed.

<Operation of content recommendation device>
Next, the operation of the content recommendation device 1 according to the present embodiment will be described.

  The content recommendation device 1 according to the present embodiment mainly performs link update processing, content recommendation processing, search response processing, and music purchase processing. Therefore, each process will be described in detail below.

≪Link update process≫
Next, details of the link update processing in the link update unit 110 of the content recommendation device 1 according to the present embodiment will be described.

  FIG. 3 is a flowchart showing the main processing for one cycle of the link update processing that is periodically operated at a predetermined time interval in the content recommendation device 1 according to the present embodiment.

  When the link update unit 110 of the content recommendation device 1 according to the present embodiment starts the link update process, the first similar node information request unit 110a reads the bibliographic related information stored in the bibliographic related information storage unit 113 ( S301). This bibliographic related information includes the music ID uniquely assigned to the music, the bibliographic information of the music, and the number of times the music has been played.

  Next, the first similar node information request unit 110a generates a search query (S302). The search query includes a list of IP addresses of nodes through which the search query has passed, a bibliographic information list generated from the bibliographic information read in step S301 and the number of reproductions, and a predetermined upper limit value of the number of nodes in the search result. . The bibliographic information list is generated as a list of bibliographic information selected by a number that does not exceed a predetermined threshold value from those with a larger number of reproductions. Note that the threshold for the number of bibliographic information and the upper limit for the number of nodes in the search result are appropriately determined in advance in consideration of the speed of the link update process, the accuracy of the content recommendation process described later, and the like.

  Next, the first similar node information request unit 110a reads the link node information stored in the link node information storage unit 115 (S303).

  Then, the first similar node information request unit 110a determines whether or not the number of link nodes in the link node information read in step S303 is smaller than a predetermined threshold (S304). If the number of link nodes is less than the threshold value, the process proceeds to step S305, and if it is equal to or greater than the threshold value, the process proceeds to step S306. Note that the threshold used here is appropriately determined within a range that does not exceed the upper limit of the number of nodes in the search result used in step S302 in consideration of the speed of link update processing and the like.

  If it is determined in step S304 that the number of link nodes is less than the threshold value, the first similar node information requesting unit 110a sends the node information that is deficient in the threshold value in step S304 from the node information server 11 via the network connection unit 104. (Step S305). The node information acquired here is randomly selected from the node information registered in the node information server 11.

  Then, the first similar node information request unit 110a adds the node information acquired from the node information server 11 to the link node information read in step S303.

  Alternatively, in the case of a configuration in which the node information server 11 is not operated, the node information obtained by the user using general-purpose information transmission means is input, and the input node information is added to the link node information read in step S303. You may do it.

  Next, the first similar node information request unit 110a transmits a similar node search request signal to each link node corresponding to the link node information read in step S303 or the link node information to which the node information is added in step S305. (Step S306). Specifically, the first similar node information request unit 110a connects to the link node having the IP address included in the link node information via the network 17 by the network connection unit 104, and requests the search for the similar link node. A similar node search request signal including the command and the search query generated in step S302 is transmitted.

  Next, the node update unit 110b receives similar node information from each link node via the network connection unit 104 as a search result corresponding to the similar node search request signal transmitted in step S306. The similar node information includes a list of node information of nodes having bibliographic information similar to the bibliographic information included in the search query and the similarity between the bibliographic information.

  Next, the first similar node information request unit 110a integrates the similar node information received from each link node in step S306 (S307). Specifically, the node information of nodes having bibliographic information similar to bibliographic information is rearranged in descending order of the similarity of bibliographic information, and the upper limit of the number of nodes used in step S302 is exceeded in descending order of bibliographic information similarity. Leave no node information and discard other search results.

  Next, the first similar node information request unit 110a causes the link node information storage unit 115 to update the node information integrated in step S307 as new link node information (S308). At this time, the node information and the similarity between the bibliographic information corresponding to the node information may be updated in the link node information storage unit 115 as new link node information.

  Next, the node information server 11 is registered with the node information of the content recommendation device 1 performing the link update process (S309).

≪Content recommendation process≫
Next, details of content recommendation processing in the content recommendation unit 109 of the content recommendation device 1 according to the present embodiment will be described.

  FIG. 4 is a flowchart showing content recommendation processing in the content recommendation unit 109 of the content recommendation device 1 according to the present embodiment.

  When the content recommendation unit 109 of the content recommendation device 1 according to the present embodiment starts the content recommendation process in response to the operation signal received from the user interface unit 102, the content list request unit 109 a links from the link node information storage unit 115. Node information is read (S401).

  Next, the content list request unit 109a connects to each link node corresponding to the IP address included in the link node information read out in step S401 via the network 17 by the network connection unit 104, and Then, a content list transmission request signal for transmitting the content list is transmitted (S402).

  Then, the content selection unit 109 b of the content recommendation unit 109 receives the content list transmitted by each link node that has received the content list transmission request signal via the network connection unit 104. The received content list includes bibliographic information of music and audio data for trial listening.

  Next, the content selection unit 109b counts the number of duplicates for each music piece from the content list received from each link node (S403). Specifically, it is determined whether or not the bibliographic information included in the content list received from each link node is the same, and if it is the same, it is determined that the music is duplicated. Then, by performing this process for each piece of music, the duplication number for each piece of music is calculated. Alternatively, if the link node information read out in step S401 includes the similarity between the bibliographic information of this node and the bibliographic information of the link node, the link node for each song from the content list received from each link node Find the total similarity of bibliographic information. Specifically, it is determined whether or not the bibliographic information included in the content list received from each link node is the same, and is included in the link node information of the link node that has transmitted each music for each music determined to be the same. By adding the bibliographic information similarities, the sum of the bibliographic information similarities of link nodes for each music is obtained.

  Next, the content selection unit 109b rearranges the songs in the descending order of the sum of the duplication numbers of the songs calculated in step S403 or the similarity of the bibliographic information of the link nodes (S404).

  Next, the content selection unit 109b reads the bibliographic information stored in the bibliographic information storage unit 113, and the bibliographic information identical to the bibliographic information of the music selected in step S404 is added to the bibliographic information of the read bibliographic information. If it is determined that there is a song, the music having the same bibliographic information is deleted from the content list (step S405).

  Next, the content selection unit 109b transmits the bibliographic information of the music selected in step S405 to the user interface unit 102, and converts the bibliographic information of the transmitted music into a list format to the user interface unit 102 as recommended content. Output (step S406).

≪Search response process≫
Next, details of the search response process in the search response unit 111 of the content recommendation device 1 according to the present embodiment will be described.

  FIG. 5 is a flowchart showing search response processing in the content recommendation device 1 according to this embodiment.

  First, the search response unit 111 of the content recommendation device 1 according to the present embodiment receives a request signal from another node via the network 17 by the network connection unit 104 (S501).

  Next, the search response unit 111 is a node having bibliographic information similar to the bibliographic information included in the search query, in which the request signal received in step S501 is a content list transmission request signal for requesting transmission of a content list. It is determined whether it is a similar node search request signal for requesting search (S502).

  Note that the search query for the similar node search request signal is generated by another node and is the same as the search query generated in step S302 of the link update process shown in FIG.

  If the search response unit 111 determines that the received request signal is a content list transmission request signal, the search response unit 111 proceeds to step S512. If the search response unit 111 determines that the received request signal is a similar node search request signal, the search response unit 111 proceeds to step S503. Transition.

  In step S502, if the search response unit 111 determines that the received request signal is a node search request signal, the second similar node information request unit 111a searches for a similar node search request signal. The number of IP addresses of the passed nodes included in the query is read, and it is determined whether or not the number of passed nodes is equal to or less than a predetermined threshold (step S503). If it is determined that the value is equal to or less than the threshold, the process proceeds to step S504. If the threshold is exceeded, the process proceeds to step S507.

  Note that the threshold value used here needs to be appropriately determined in advance in consideration of the range in which the similar node search request signal is transmitted on the P2P network.

  Next, when the second similar node information request unit 111a determines in step S503 that the number of passed nodes is equal to or less than a predetermined threshold, the second similar node information request unit 111a receives The IP address of the content recommendation device 1 is added to the list of IP addresses of the search query included in the similar node search request signal (step S504).

  Next, the second similar node information request unit 111a reads the link node information stored in the link node information storage unit 115 (S505).

  Then, the second similar node information request unit 111a includes a command for requesting a search for a similar link node for each link node corresponding to the link node information read in step S505, and the search query changed in step S504. A similar node search request signal is transmitted via the network connection unit 104 (S506). However, in order to avoid looping the transmission of the similar node search request signal, the link node whose IP address is included in the IP address list of the search query is excluded from the transmission target of the similar node search request signal.

  Then, the node transmission unit 111b of the search response unit 111 receives the similar node information as a search result for the similar node search request signal transmitted by each link node via the network connection unit 104. The similar node information includes node information of a node having bibliographic information similar to the bibliographic information included in the search query and the similarity between the bibliographic information.

  Next, the node transmission unit 111b reads the bibliographic information stored in the bibliographic information storage unit 113 (step S507). Then, from the read bibliographic information, the number of bibliographic information that does not exceed a predetermined threshold is selected from the one with the larger number of reproductions.

  Then, the degree of similarity between the bibliographic information included in the bibliographic information selected in step S507 and the bibliographic information included in the search query of the received similar node search request signal is calculated (S508). Specifically, the bibliographic information included in the selected bibliographic information and the bibliographic information included in the received similar node search request signal search query are used to calculate the number of the same bibliographic information. Let the number of information be the similarity.

  In this way, the node transmission unit 111b of the search response unit 111 can calculate the similarity between the bibliographic information of the node that transmitted the similar node search request signal and the bibliographic information of the content recommendation device 1.

  Next, the node transmitting unit 111b adds the node information of the content recommendation device 1 as one piece of similar node information to the similar node information received in step S505 or empty similar node information if step S505 is not performed. The similarities of the bibliographic information obtained in S508 are added in association with each other, and are generated as all similar node information (S509).

  Then, the node transmission unit 111b selects all similar node information (S510). Specifically, the node information included in all the similar node information generated in step S509 is rearranged in descending order of the similarity of the bibliographic information, and the nodes of the search results included in the search query from the higher bibliographic information similarity The number of pieces of node information that does not exceed the upper limit of the number and the similarity of bibliographic information corresponding to the node information are selected as similar node information. Other similar node information is discarded.

  Next, the node transmission unit 111b returns the similar node information newly generated in step S510 to the node that transmitted the similar node search request signal via the network 17 (S511).

  On the other hand, when the search response unit 111 determines in step S502 that the content list transmission request signal is received, the content list response unit 111c reads the bibliographic information from the bibliographic information storage unit 113 and reproduces the read bibliographic information. The number of bibliographic related information that does not exceed a predetermined threshold value is selected from the larger number of times (step S512).

  Then, in step S512, the content list response unit 111c reads the audition audio data of the music corresponding to the music ID of the selected bibliographic related information from the audition audio data storage unit 114 (step S513).

  Then, the content list response unit 111c transmits the bibliographic information of the bibliographic information selected in step S512 and the audio data for audition read in step S513 to the node that transmitted the content list transmission request signal. (S514).

≪Music purchase processing≫
Next, details of the music purchase processing in the music purchase processing unit 108 of the content recommendation device 1 according to the present embodiment will be described.

  FIG. 6 is a flowchart showing the music purchase processing in the music purchase processing unit 108 of the content recommendation device 1 according to the present embodiment.

  First, when the user performs a selection operation of a desired song from the bibliographic information of the recommended content selected by the content recommendation unit 109 and output by the user interface unit 102, the song purchase processing unit 108 selects the user of the song. The signal is received, and the bibliographic information of the music corresponding to the user selection signal is selected from the bibliographic information of the recommended content (step S601).

  Next, the music purchase processing unit 108 connects to the content providers 18 and 19 via the network 17 from the network connection unit 104, and searches for the music selected in step 601 using the communication interface for each of the content providers 18 and 19 ( Step S602). Specifically, a request signal for inquiring whether or not there is bibliographic information that matches the bibliographic information of the music selected in step S601 is transmitted to the content providers 18 and 19, and the content provider 18 that has transmitted a signal indicating that it has bibliographic information is transmitted. , 19 lists.

  Here, it is assumed that the content recommendation device 1 according to the present embodiment corresponds to a plurality of content providers 18 and 19. Therefore, it is assumed that the content providers 18 and 19 also have a search function based on bibliographic information. However, since the response interface may differ depending on the content providers 18 and 19, the request signal for inquiring whether there is bibliographic information that matches the bibliographic information of the content selected in step S601 is adapted to each content provider 18 and 19. It will be done at the interface.

  Then, the music purchase processing unit 108 outputs the content provider list generated in step S602 to the user interface unit (step S603).

  Next, when the user performs a selection operation of a desired content provider from the content provider list, the music purchase processing unit 108 receives the user selection signal of the content provider and uses it from the content provider corresponding to the user selection signal. The purchase process of the music piece desired by the user is executed (step S605).

  As described above, mainly by performing content recommendation processing, link update processing, and search support processing, collaborative filtering, that is, music recommendation can be performed based on the content list of link nodes with similar bibliographic information. it can. Then, by performing the music purchase process, the user can purchase music of recommended content recommended from a plurality of content providers.

  As a result, content recommendation can be performed appropriately for each user, and further, an increase in cost due to the use of a server having hardware and a network environment that can withstand a considerable load can be prevented.

  The embodiment described above can also be realized by executing a content recommendation program installed in a computer. The content recommendation program may be read from a recording medium storing the content recommendation program and executed by the CPU to constitute a content recommendation device, or may be transmitted via a communication network. The content recommendation device may be configured by being installed and executed by the CPU.

  In order to increase the processing speed of the link update process in the link update unit 110, the second similar node information request unit 111a in the search response unit 111 may be omitted. In this case, since the similar node search request signal does not propagate on the P2P network, it is considered that the accuracy of the link update process is lower than that of the configuration having the second similar node information request unit 111a. Increases speed and reduces the load on the network.

It is the block diagram which showed the structure of the content recommendation apparatus which is this embodiment. It is the block diagram which showed the structure of the content recommendation system to which the content recommendation apparatus which is this embodiment is applied. It is the flowchart which showed the link update process in the content recommendation apparatus which is this embodiment. It is the flowchart which showed the content recommendation process in the content recommendation apparatus which is this embodiment. It is the flowchart which showed the search response process in the content recommendation apparatus which is this embodiment. It is the flowchart which showed the music purchase process in the content recommendation apparatus which is this embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Content recommendation apparatus 2 ... Audio | voice output part 10 ... Content recommendation system 11 ... Node information server 12-16 ... Content recommendation apparatus 17 ... Network 18, 19 ... Content provider 101 ... Audio data input part 102 ... User interface part 103 ... Audio data decoding unit (content output means)
DESCRIPTION OF SYMBOLS 104 ... Network connection part 105 ... Music taking part 106 ... Audio data encoding part 107 ... Music reproduction part 108 ... Music purchase process part 109 ... Content recommendation part 109a ... Content list request means 109b ... Content selection means 109c ... Sample output control Means 110 ... Link update unit 110a ... First similar node information request means 110b ... Node update means 111 ... Search response section 111a ... Second similar node information request means 111b ... Node transmission means 111c ... Content list response means 112 ... Code Audio data storage unit 113... Bibliographic information storage unit 114. Audition audio data storage unit (sample content data storage unit)
115 ... Link node information storage unit (node information storage unit)
116 ... Audio data generation unit for trial listening

Claims (4)

A content recommendation device that is provided in each of a plurality of playback devices capable of playing back content and is peer-to-peer connected via a network,
Bibliographic information storage means for storing bibliographic information related to the content;
Link node information storage means for storing link node information for specifying other content recommendation devices other than the content recommendation device;
Reading each of the stored bibliographic information and link node information, and specifying a content recommendation device having bibliographic information similar to the read bibliographic information for other content recommendation devices corresponding to the read link node information First similar node information requesting means for requesting transmission of similar node information including similarity between the bibliographic information possessed by the identified content recommendation device and the read bibliographic information,
Node update means for receiving similar node information transmitted from the other content recommendation device, and updating link node information stored in the link node information storage means based on the received similar node information;
Link node information stored in the link node information storage means when the transmission of the similar node information is requested from the other content recommendation device, and the number of devices through which the transmission request has passed is equal to or less than a predetermined threshold. Second similar node information requesting means for requesting transmission of similar node information to other content recommendation devices corresponding to the read link node information,
The similar node information requested to be transmitted by the second similar node information request unit is received, the bibliographic information stored in the bibliographic information storage unit is read out, and the read bibliographic information is included in the similar node information. If the similarity is similar, the device information and the similarity between the read bibliographic information and the bibliographic information included in the similar node information are added to the similar node information. Node transmitting means for generating a number of similar node information that does not exceed the upper limit as new similar node information, and transmitting the generated new similar node information to the content recommendation device that has requested transmission of the similar node information;
The link node information updated by the node update unit is read from the link node information storage unit, and another content recommendation device corresponding to the read link node information is provided with a playback device provided with the other content recommendation device. Content list requesting means for requesting transmission of a content list which is a list of the contents having;
Content selection means for receiving a content list transmitted from the other content recommendation device and selecting a predetermined number of content as recommended content based on the received content list;
A content recommendation device comprising: output means for outputting the selected recommended content to the playback device. Sample content data storage means for storing sample data for trial viewing corresponding to the content;
When the transmission of the content list is requested, the bibliographic information stored in the bibliographic information storage unit is read, a predetermined number of bibliographic information is selected from the read bibliographic information, and the selected bibliographic information is stored in the selected bibliographic information. Content response means for reading corresponding sample data from the sample content data storage means, and sending the read sample data and the selected bibliographic information as a content list to a content recommendation device that has requested transmission of the content list;
Content received by the content selection means when sample content corresponding to the content indicated by the received content selection signal is received when a content selection signal for selecting content from the recommended content is received by a user operation The content recommendation device according to claim 1, further comprising: sample output control means for selecting from a list and outputting to a content output means. A content recommendation device provided in each of a plurality of reproduction devices capable of reproducing content and connected peer-to-peer via a network, and bibliographic information storage means for storing bibliographic information related to the content, and other than the content recommendation device A content recommendation program for causing the content recommendation apparatus to include link node information storage means for storing link node information for specifying another content recommendation apparatus,
Read the bibliographic information and link node information stored from the bibliographic information storage unit and the link node information storage unit, respectively, and send the read bibliographic information to another content recommendation device corresponding to the read link node information. Requesting transmission of similar node information including the degree of similarity between the device information for specifying the content recommendation device having bibliographic information similar to the content and the bibliographic information of the specified content recommendation device and the read bibliographic information. 1 similar node information request step;
A node update step of receiving similar node information transmitted from the other content recommendation device, and updating link node information stored in the link node information storage unit based on the received similar node information;
Link node information stored in the link node information storage means when the transmission of the similar node information is requested from the other content recommendation device, and the number of devices through which the transmission request has passed is equal to or less than a predetermined threshold. A second similar node information requesting step for requesting transmission of similar node information to another content recommendation device corresponding to the read link node information,
The similar node information requested for transmission in the second similar node information requesting step is received, the bibliographic information stored in the bibliographic information storage means is read, and the bibliographic information in which the read bibliographic information is included in the similar node information If the similarity is similar, the device information and the similarity between the read bibliographic information and the bibliographic information included in the similar node information are added to the similar node information. A node transmission step of generating a number of similar node information that does not exceed the upper limit as new similar node information, and transmitting the generated new similar node information to the content recommendation device that has requested transmission of the similar node information;
The link node information updated by the node update step is read from the link node information storage means, and a playback device in which the other content recommendation device is provided in another content recommendation device corresponding to the read link node information. A content list requesting step for requesting transmission of a content list which is a list of the contents having;
A content selection step of receiving a content list transmitted from the other content recommendation device and selecting a predetermined number of content as recommended content based on the received content list;
A content recommendation program for causing the content recommendation device to execute an output step of outputting the selected recommended content to the playback device. The content recommendation device further includes:
Sample content data storage means for storing sample data for trial viewing corresponding to the content,
When the transmission of the content list is requested, the bibliographic information stored in the bibliographic information storage unit is read, a predetermined number of bibliographic information is selected from the read bibliographic information, and the selected bibliographic information is selected. A content response step of reading the sample data corresponding to the sample content data storage means and transmitting the read sample data and the selected bibliographic information as a content list to the content recommendation device that has requested transmission of the content list; ,
Content received in the content selection step when sample data corresponding to the content indicated by the received content selection signal is received when a content selection signal for selecting content from the recommended content is received by a user operation The content recommendation program according to claim 3 , further causing the content recommendation device to execute a sample output control step of selecting and outputting from the list.

Images