JP2013149140A - Information processing device and information processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To recommend an item from a different service on the basis of action information on a user.SOLUTION: An information processing device includes: a transmission/reception unit that stores in a memory action history information indicating a history of operation performed on a terminal by a user for receiving a service; item information that includes plural candidates for an item to be recommended to the user and an attribute value of each candidate; weighting factor information that includes a weighting factor indicating the importance of an attribute assigned to each of plural groups in which users are to be placed according to histories; a group determination unit that determines a group in which the user is placed on the basis of the action history information; and a degree-of-recommendation calculation unit that calculates similarity in attribute values between items on the basis of the weighting factor assigned to the determined group and the item information. The degree-of-recommendation calculation unit calculates the degree of recommendation of each candidate according to the similarity in attribute values between items and determines an item to be recommended to the user from the candidates according to the calculated degree of recommendation.

Description

  The present invention relates to an information processing apparatus, and more particularly to an information processing apparatus that recommends an item to a user.

  2. Description of the Related Art In recent years, research and development have been actively conducted on a recommendation technique in which a computer extracts an item that matches a user's preference from a large number of items and recommends the extracted item to the user.

  For example, a recent digital television has a function of extracting features of a program viewed or recorded by a user and recommending a future program having features similar to the extracted features to the user.

  Such a system that provides recommendation technology such as digital television has attribute values such as “genre”, “channel”, or “time zone” for each program, and each attribute of each program that is viewed and recorded. User preference information is acquired by vectorizing values. On the other hand, the attributes of future programs to be recommended are also vectorized. And the system which provides the recommendation technique calculates the similarity between the vector of each attribute, and recommends the future program with a high similarity to a user.

  When calculating the similarity of attribute vectors, a method such as calculating the Euclidean distance or inner product is used. In either method, a weighting factor indicating which attribute is important is set, and each attribute is calculated. A method of reflecting a weighting factor in the calculation is often used.

  Conventionally, a common fixed value that does not depend on an individual has been used for this weighting coefficient. However, the attributes that are considered important for each user are different. For this reason, a technique for recommending an item that matches the user's preference by changing the weighting factor for each user has been proposed.

  As a technique for recommending items that match the user's preference, it is determined that a user's operation of viewing or recording content is a positive operation, and an operation of deleting content is a negative operation, and the determination result is similar to each attribute. A technique for determining a weighting factor of an attribute of content based on multiple regression analysis using degrees has been proposed (see, for example, Patent Document 1).

  On the other hand, in recent years, in a specific service, user preference information is extracted from user behavior or operation, and based on the extracted preference information, not only the items of the same service are recommended, but also preference information is provided. Research and development of recommendation technology that recommends items of services different from the extracted services are underway.

  Such recommendation technology includes, for example, a recommendation technology that recommends healthy goods when a user who often watches a health program on a television uses an internet supermarket. This system for providing recommendation technology extracts user preference information when providing a service called TV, and provides items (for example, health goods) based on the extracted preference information when providing a service called Internet supermarket. ) Is recommended.

JP 2010-061601 A

  As described above, according to the technology such as Patent Document 1, the user's preference is determined based on the information on the user's behavior or operation for each item in the target service and the attribute value assigned to each item. The weighting factor of each attribute value reflected is determined. By using the weighting factor determined in this way, it is possible to recommend an appropriate item to the user.

  However, in the technique such as Patent Document 1, the attribute of the item used for extracting the preference information and the attribute of the item to be recommended are the same. That is, it is assumed that preference information is extracted in the same service and items are recommended.

  For this reason, when the service for extracting user preference information is different from the service for recommending items to the user, it is difficult to determine the weighting factor of the item attribute that appropriately reflects the user preference. As a result, there is a problem that it is difficult to recommend an optimal item to the user when crossing between services.

  Further, in the technology such as Patent Document 1, when presenting a recommended item to the user, it is not considered to present the reason for the recommendation, and the user can determine why the item was recommended. Since it cannot be grasped, there is a problem that an uncomfortable feeling is obtained.

  An object of the present invention is to provide an information processing apparatus that recommends an item that matches a user's state and preference from a large number of items, and recommends an optimal item to a user even when a plurality of different services are used. It is an object of the present invention to provide an information processing apparatus that makes it possible to Moreover, when presenting the recommended item to a user, it is a subject to provide the information processing apparatus which a user does not get discomfort by also showing the reason for recommendation together.

  A typical example of the present invention is as follows. That is, an information processing apparatus that receives data from a terminal that provides a service to a user, the information processing apparatus including a processor, a memory, and an interface, and the user operating the terminal to receive the service When acquiring action history information indicating a history, an item including a transmission / reception unit that stores the acquired action history information in the memory, a plurality of candidates for items recommended to the user, and an attribute value of each candidate A group for classifying the user based on information, weight coefficient information including a weight coefficient indicating the importance of the attribute assigned to a plurality of groups in which the user is classified according to the history, and the action history information A group deciding unit for deciding on the basis of the weight coefficient information and the item information. A recommendation level calculation unit that calculates a distance between the assigned weighting coefficient and the attribute value of each candidate, and the recommendation level calculation unit calculates the recommendation level of each candidate according to the calculated distance. An item to be recommended to the user is determined from the candidates according to the calculated recommendation degree.

  According to an embodiment of the present invention, items of different services can be recommended from behavior information acquired in one service.

It is explanatory drawing which shows the physical structure of the information processing apparatus of embodiment of this invention. It is a block diagram which shows the hardware constitutions of the server of embodiment of this invention. It is a block diagram which shows the hardware constitutions of the portable terminal of embodiment of this invention. It is a block diagram which shows the function which the server of embodiment of this invention has. It is a block diagram which shows the function which the portable terminal of embodiment of this invention has. It is explanatory drawing which shows personal information DB of embodiment of this invention. It is explanatory drawing which shows action history DB of embodiment of this invention. It is explanatory drawing which shows item DB of embodiment of this invention. It is explanatory drawing which shows weighting coefficient DB of embodiment of this invention. It is a flowchart which shows the process which determines the recommendation degree of embodiment of this invention. It is explanatory drawing which shows the recommendation degree of embodiment of this invention. It is explanatory drawing which shows the display screen of the recommendation result data of embodiment of this invention. It is explanatory drawing which shows the display screen of the recommendation result data of embodiment of this invention. It is explanatory drawing which shows item evaluation DB of embodiment of this invention. It is explanatory drawing which shows the calculation method of the recommendation degree of embodiment of this invention.

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

  In addition, in this embodiment, about the component which has the substantially the same function and structure, the overlapping description is abbreviate | omitted by attaching | subjecting the same code | symbol.

  First, FIG. 1 shows an example of the system configuration of the information processing apparatus of this embodiment.

  FIG. 1 is an explanatory diagram showing a physical configuration of the information processing system according to the embodiment of this invention.

  The information processing system of this embodiment is implemented by a single unit such as a server, a television, a PC, a recorder, a mobile phone, and a mobile terminal, or a combination of these units. The information processing system illustrated in FIG. 1 is a system including a server and a mobile terminal. The information processing system in the following description is configured by a server and a mobile terminal, but the information processing system of the present invention is not limited to this.

  The information processing system according to the present embodiment illustrated in FIG. 1 includes a server 101, a network 102, at least one base station 103, and at least one portable terminal 104.

  The server 101 is an information processing apparatus such as a computer system having a network communication function. The server 101 processes the request transmitted from the mobile terminal 104 and transmits a response to the mobile terminal 104.

  The mobile terminal 104 is a device such as a PDA (Personal Digital Assistance) or a mobile phone having a network communication function. The portable terminal 104 connects to the server 101 via the base station 103 and the network 102 according to the operation by the user.

  The mobile terminal 104 transmits a request to the server 101 in order to receive the service desired by the user from the server 101. And a service is provided to a user by processing the response corresponding to the request transmitted from the server 101.

  For example, the mobile terminal 104 is a terminal device that provides a service to a user, such as displaying information provided by the service on a screen provided in the mobile terminal 104.

  The network 102 is a network such as the Internet or an intranet, for example. Here, the intranet includes a cellular phone network, a LAN (Local Area Network), a WAN (Wide Area Network), and the like.

  The base station 103 is, for example, a terminal device in a mobile phone network or an access point such as a wireless LAN. The server 101 and the mobile terminal 104 transmit / receive data including a request or a response via the network 102 and the base station 103.

  The mobile terminal 104 shown in FIG. 1 stores, in the server 101, information related to actions performed by the user using the mobile terminal 104, information related to user operations on the mobile terminal 104, and setting information set on the mobile terminal 104. Send.

  In the present embodiment, the action performed by the user using the portable terminal 104 is an action performed by the user for receiving a service via the portable terminal 104, such as shaking the portable terminal 104 or holding the portable terminal 104 over a predetermined device. All of that is included. In the following description, it is described that the user's operation on the mobile terminal 104 includes an action that the user performs using the mobile terminal 104.

  The server 101 illustrated in FIG. 1 selects items that are candidates for recommendation to the user from a large number of items held by the server 101 based on information received from the mobile terminal 104. Then, the server 101 transmits information regarding the selected item and the reason for selection to the mobile terminal 104.

  Then, the mobile terminal 104 shown in FIG. 1 displays the information related to the item transmitted from the server 101 and the reason for selection on the screen included in the mobile terminal 104.

  Next, an example of the hardware configuration of the server 101 and the mobile terminal 104 according to the embodiment of this invention is shown.

  FIG. 2 is a block diagram illustrating a hardware configuration of the server 101 according to the embodiment of this invention.

  The server 101 includes at least a communication device 201, a processor 202, a storage device 203, and a secondary storage device 204. Each device included in the server 101 is connected by a bus 205. Therefore, data can be transmitted and received between the devices.

  The communication device 201 is a network card unit such as a LAN card, for example, and is a device for connecting the server 101 to the network 102.

  The processor 202 is, for example, a device such as a CPU, and may be configured mainly with a microprocessor. The processor 202 is a device that executes a program stored in the storage device 203 or the secondary storage device 204 and controls processing in the server 101.

  The storage device 203 is, for example, a random access memory (RAM). The storage device 203 is a device that stores programs executed by the processor 202, data received from the mobile terminal 104, and the like, and stores data processed by the server 101.

  The secondary storage device 204 is, for example, a device such as a hard disk device or a nonvolatile memory. The secondary storage device 204 stores search target information and a database including the search target information. Further, the secondary storage device 204 stores various data generated in a recommendation degree calculation process described later, and a database including the data.

  FIG. 3 is a block diagram illustrating a hardware configuration of the mobile terminal 104 according to the embodiment of this invention.

  The portable terminal 104 includes at least a communication device 301, a processor 302, an input device 303, a storage device 304, and a display device 305. Note that the devices included in the mobile terminal 104 are connected by a bus 306, and data can be transmitted and received between the devices.

  The communication device 301 is implemented by, for example, a module for connecting a mobile phone and a mobile phone network and a driver thereof, a module for connecting a mobile phone and a wireless LAN, a driver thereof, and the like. The communication device 301 connects the mobile terminal 104 to the network 102.

  The processor 302 is configured mainly with a microprocessor, executes a program stored in the storage device 304, and controls processing in the portable terminal 104.

  The input device 303 is mounted by a touch panel in a PDA or the like and its driver. The user can input a service request to the mobile terminal 104 using the input device 303. In particular, the user in the present embodiment inputs an operation for receiving services such as viewing of video content and purchase of products using the input device 303.

  The storage device 304 is implemented by, for example, a random access memory (RAM). The storage device 304 stores a program executed by the processor 302, data received from the server 101 and processed in the portable terminal 104, and the like.

  The display device 305 is mounted by, for example, a liquid crystal panel provided in a PDA or the like and a driver thereof. The display device 305 displays data that the mobile terminal 104 has. The display device 305 of the present embodiment displays information related to the item selected as a candidate from the server 101 and the selection reason.

  FIG. 4 is a block diagram illustrating functions of the server 101 according to the embodiment of this invention.

  In the present embodiment, all functions of the server 101 are described as being implemented by a software program executed by the processor 202, but some or all of the functions may be implemented by hardware.

  In addition to the functions shown in FIG. 4, the server 101 includes an OS (Operating System) that performs activation and operation management of each function, a control program, and the like. However, hereinafter, the processing of the function shown in FIG.

  The server 101 includes a request response transmission / reception unit 401, a request response processing unit 402, a recommendation calculation control unit 403, and a recommendation device 404.

  The request response transmission / reception unit 401 receives a request transmitted from the mobile terminal 104 to the server 101 using the communication device 201 and transmits a response transmitted from the server 101 to the mobile terminal using the communication device 201. Have

  The request response processing unit 402 has a function of processing data transmitted / received by the request response transmitting / receiving unit 401, that is, data transmitted / received between the mobile terminal 104 and the server 101.

  Specifically, the request response processing unit 402 acquires the personal information acquired by the mobile terminal 104 via the request response transmission / reception unit 401. Then, the request response processing unit 402 registers the acquired personal information in the personal information DB 405 in the recommendation device 404.

  Further, the request response processing unit 402 acquires the action history information acquired by the mobile terminal 104 via the request response transmission / reception unit 401. Then, the request response processing unit 402 registers the acquired action history information in the action history DB 406 in the recommendation device 404.

  Further, the request response processing unit 402 receives a request for obtaining recommendation result data from the portable terminal 104 via the request response transmission / reception unit 401. When receiving the recommendation result data acquisition request, the request response processing unit 402 acquires the recommendation result data 414 in the recommendation device 404 and acquires the recommendation result acquired in the portable terminal 104 via the request response transmission / reception unit 401. Send data.

  The recommendation calculation control unit 403 has a function of instructing the recommendation device 404 to start processing for calculating a recommendation level. Details of the process of calculating the recommendation level will be described later.

  Further, the process of calculating the recommendation level depends on the system design and may be started at any timing. For example, it may be started once a day at a predetermined time, or may be started when an execution request from a user is received via the request response transmission / reception unit 401.

  The recommendation device 404 includes a personal information DB 405, an action history DB 406, an item DB 407, an attribute group determination unit 408, a weight coefficient DB 409, a weight coefficient determination unit 410, an item evaluation DB 411, an item evaluation generation unit 412, A recommendation calculation unit 413 and recommendation result data 414 are included.

  The personal information DB 405 is a database that stores personal information indicating information related to individual users. The personal information of the present embodiment is data indicating, for example, the user's “age”, “sex”, “address”, “height”, “weight”, and the like.

  The personal information stored in the personal information DB 405 is data determined or inferred by the mobile terminal 104 based on data input by the user via the screen of the mobile terminal 104, a history of operations on the mobile terminal 104 by the user, or the like. is there. The portable terminal 104 transmits personal information to the recommendation device 404 via the request response transmission / reception unit 401 and the request response processing unit 402.

  The behavior history DB 406 is a database that stores a behavior history indicating a history of operations performed on the mobile terminal 104 by the user. The action history of the present embodiment is data indicating actions such as purchase of products and viewing of video content, for example.

  The action history of the present embodiment is data determined or inferred by the mobile terminal 104 from an input operation performed by the user via the screen of the mobile terminal 104. The portable terminal 104 transmits the action history to the recommendation device 404 via the request response transmission / reception unit 401 and the request response processing unit 402.

  The item DB 407 is a database that stores item information indicating items to be presented to the user. Item information is stored in the item DB 407 by the administrator who provides the service. The item DB 407 may hold data for various services such as a net supermarket or health care.

  In the present embodiment, items of the same service have the same attribute, and each item is assigned a value in advance for each attribute. The item DB 407 stores attribute values assigned to items.

  The attribute group determination unit 408 is activated by the recommendation calculation control unit 403, and acquires personal information and an action history regarding a user who recommends an item from the personal information DB 405 and the action history DB 406. And based on the acquired personal information and action history, a user's attribute group is determined by the process mentioned later.

  The user attribute group includes, for example, “health-oriented”, “thrifter”, and “gourmet-oriented”. If personal information cannot be acquired from the viewpoint of personal privacy, the attribute group determination unit 408 acquires the behavior history from the behavior history DB 406 without acquiring the personal information from the personal information DB 405, and only acquires the acquired behavior history. Based on this, an attribute group of the user may be determined.

  The weighting coefficient DB 409 is a database that stores, for each attribute group, a weighting coefficient set indicating a collection of values (weighting coefficients) indicating the importance of item attributes. The weighting coefficient DB 409 may include importance levels of attributes that are different for each service.

  The weight coefficient determination unit 410 is activated by the recommendation calculation control unit 403. The weighting factor determination unit 410 acquires a weighting factor set in the target service from the weighting factor DB 409 using the user attribute group determined by the attribute group determination unit 408 as a key. Then, the recommendation calculation unit 413 is notified of the acquired weight coefficient set.

  The item evaluation DB 411 is a database that stores an evaluation value indicating whether or not each user regards each item as important. It is updated by an item evaluation generation unit 412 described later.

  The item evaluation generation unit 412 is activated by the recommendation calculation control unit 403. The item evaluation generation unit 412 acquires an action history related to the user from the action history DB 406. Then, based on the acquired action history, the evaluation value of the user for each item stored in the item DB 407 is calculated, and the calculated evaluation value is stored in the item evaluation DB 411.

  The recommendation calculation unit 413 is activated by the recommendation calculation control unit 403. The recommendation calculation unit 413 includes information on the attribute value of each item stored in the item evaluation DB 411, the weight coefficient determined by the weight coefficient determination unit 410, and the evaluation values of all users stored in the item evaluation DB 411. Is entered. Then, the recommendation calculation unit 413 calculates the recommendation level of each item based on the input value, and generates recommendation result data 414 according to the calculated recommendation level.

  Note that if the user's evaluation value cannot be acquired, the recommendation device 404 may not have the item evaluation DB 411 and the item evaluation generation unit 412. In this case, the recommendation calculation unit 413 receives only the information regarding the attribute value of each item stored in the item evaluation DB 411 and the weighting factor determined by the weighting factor determination unit 410, and recommends based on the input value. The degree may be calculated.

  The recommendation result data 414 stores information related to items recommended to the user. For example, the recommendation level for each item stored in the item DB 407 may be stored, or information on several items may be stored in descending order of recommendation level.

  Further, the recommendation result data 414 may include information on the attribute group determined by the attribute group determination unit 408 as information on the reason for recommendation, and has the highest value among the weighting factors determined by the weighting factor determination unit 410. It may also contain information about (important) attributes.

  The recommendation result data 414 generated by the recommendation calculation unit 413 is transmitted to the mobile terminal 104 via the request response processing unit 402 and the request response transmission / reception unit 401.

  The personal information DB 405, the action history DB 406, the item DB 407, the weighting coefficient DB 409, the item evaluation DB 411, and the recommendation result data 414 are stored in the secondary storage device 204. In addition, the data is temporarily read into the storage device 203 as necessary.

  FIG. 5 is a block diagram illustrating functions of the mobile terminal 104 according to the embodiment of this invention.

  In the present embodiment, all functions of the mobile terminal 104 are described as being implemented by a software program executed by the processor 302, but some or all functions of the mobile terminal 104 are implemented by hardware. Also good.

  In addition to the functions illustrated in FIG. 5, the mobile terminal 104 includes an OS (Operating System) that performs activation and operation management of each function, a control program, and the like. However, hereinafter, the processing of the functions shown in FIG.

  The portable terminal 104 includes an input unit 501, a request response transmission / reception unit 502, a request response processing unit 503, and a display unit 504.

  The input unit 501 is a function for the user to input instructions indicating personal information registration, product purchase, content viewing, and the like. The input unit 501 acquires information on the information displayed on the display device 305 by the display unit 504, which will be described later, by the user operating the input device 303. Thereby, the input unit 501 can acquire the user's personal information and the history of operations performed on the mobile terminal 104 by the user.

  The request response transmission / reception unit 502 transmits the user's personal information acquired by the input unit 501 and information indicating the history of operations performed on the mobile terminal 104 by the user to the server 101 via the communication device 301. In addition, a request for recommending an item is transmitted to the server 101 via the communication device 301. Further, the recommendation result data 414 generated by the server 101 is received via the communication device 301.

  Based on the recommendation result data 414 received by the request response transmission / reception unit 502, the request response processing unit 503 displays an output screen for displaying information regarding the item recommended from the server 101 and the reason for recommendation on the display device 305. Generate.

  The display unit 504 displays the output screen generated by the request response processing unit 503 on the display device 305.

  When the portable terminal 104 is not a portable terminal but a fixed device such as a PC, the display unit 504 may cause the device such as a printer to print the output screen generated by the request response processing unit 503.

  The mobile terminal 104 of the present embodiment executes applications for providing various services such as “net super”, “TV”, and “health care”.

  FIG. 6 is an explanatory diagram showing the personal information DB 405 according to the embodiment of this invention.

  The personal information DB 405 includes an area 601 and areas 602 to 607 that store information related to personal information. The personal information DB 405 illustrated in FIG. 6 includes a row 611 and a row 612.

  The content of information stored in each area of the personal information DB 405 is determined in advance by the administrator. Further, the information stored in the personal information DB 405 is not limited to the content shown in FIG.

  An area 601 is an identifier for uniquely identifying a user, and is an ID unique to each user. The request response processing unit 402 of the server 101 acquires from the portable terminal 104 that the number of new users using the portable terminal 104 has increased, and assigns a new identifier to be stored in the area 601 to the new user.

  The values in the areas 602 to 608 are values generated based on information registered by the user in the mobile terminal 104. For this reason, each cell included in the personal information DB 405 may include a cell that does not include a value.

  An area 602 indicates the “name” of the user. An area 603 indicates the “age” of the user. An area 604 indicates the “gender” of the user.

  An area 605 shows the user's “hobby”. An area 606 shows the “blood pressure value” of the user. An area 607 shows the “medical history” of the user.

  The value included in the personal information DB 405 may be a numerical value or a character string.

  In the present embodiment, the user registers his / her personal information via a registration screen displayed on the display device 305 of the mobile terminal 104. The request response processing unit 503 of the portable terminal 104 receives the input personal information via the input unit 501 and includes the personal information in a signal transmitted to the server 101.

  Then, the request response processing unit 503 transmits a signal including personal information to the server 101 via the request response transmission / reception unit 502. The request response processing unit 402 of the server 101 receives the request response transmission / reception unit 401 and registers the received personal information in the personal information DB 405.

  Here, the request response processing unit 402 of the server 101 may store not only the personal information transmitted from the portable terminal 104 but also information estimated from the transmitted personal information in the personal information DB 405.

  For example, when the user inputs a blood pressure value to the mobile terminal 104, the request response processing unit 402 not only stores the blood pressure value in the area 606 corresponding to the user in the personal information DB 405, but also stores the medical history based on the blood pressure value. You may judge. Specifically, when the blood pressure value is higher than a predetermined value, the request response processing unit 402 may determine that the user's medical history is high blood pressure, and store the determined medical history in the area 607 corresponding to the user. .

  FIG. 7 is an explanatory diagram illustrating the action history DB 406 according to the embodiment of this invention.

  The action history DB 406 shown in FIG. 7 shows an action history when the mobile terminal 104 provides services of “net super”, “TV”, “video content”, and “recipe” to a user via a dedicated application or the like. Show.

  Lines 711 to 717 show the action history of each user. The action history DB 406 includes areas 701 to 705. Further, the action history DB 406 illustrated in FIG. 7 includes rows 711 to 717.

  Values stored in the areas 701 and 702 are defined in advance by the administrator, and the values stored in the areas 703 to 705 are stored based on information related to user operations.

  An area 701 indicates a user ID and corresponds to the area 601. An area 702 indicates the type of service provided to the user.

  An area 703 indicates the time when the user operates the mobile terminal 104. An area 704 indicates the type of operation performed by the user. An area 705 indicates a target of an operation performed by the user.

  When the user performs a specific operation on a specific target provided by the service, the request response processing unit 503 of the mobile terminal 104 acquires an action history related to the user. Then, the acquired action history is transmitted to the server 101 via the request response transmission / reception unit 502.

  The request response processing unit 402 of the server 101 receives the behavior history via the request response transmission / reception unit 401, and stores the received behavior history in the behavior history DB 406.

  For example, in the rows 711 and 712 shown in FIG. 7, a user whose identifier is “1” (hereinafter referred to as user 1) purchases a vegetable called “AAA” and calls it “BBB” in the service of the net supermarket. Indicates that you have browsed information on health goods.

  A row 713 indicates that the user 1 has recorded a health program called “CCC” in the TV service. A row 714 indicates that a user whose identifier is “2” (hereinafter, user 2) has watched a cooking program called “DDD”.

  Lines 715 and 716 indicate that the user 2 downloaded a variety content called “EEE” and viewed a variety content called “EEE” in the video content service.

  A row 717 indicates that the user 2 browsed a recipe for cooking called “FFF” in the recipe service.

  FIG. 8 is an explanatory diagram illustrating the item DB 407 according to the embodiment of this invention.

  Item DB407 contains the information regarding the item which calculates a recommendation degree. The item DB 407 illustrated in FIG. 8 includes information regarding items when the service is “recipe”.

  The item DB 407 includes areas 801 to 808. The item DB 407 illustrated in FIG. 8 includes rows 811 to 815.

  The value stored in each area of the item DB 407 is predetermined by the administrator. However, the values of the areas 802 to 808 are not limited to the values shown in FIG.

  An area 801 shows an identifier for uniquely identifying an item. An area 802 shows the name of the item. In the area 802 shown in FIG. 8, the name of the dish is stored. Areas 803 to 808 indicate item attributes.

  An area 803 shown in FIG. 8 shows a menu to which the item belongs. An area 804 shows the genre to which the item belongs. Region 805 shows the calories of the item.

  A region 806 shows the amount of salt contained in the item. A region 807 indicates the amount of protein contained in the item. A region 808 shows cooking time until an item is made.

  Each cell of the item DB 407 may include a numerical value or a character string.

  Lines 811 to 815 indicate attribute values of the items.

  The item DB 407 illustrated in FIG. 8 includes items provided by the “recipe” service, but may include items provided by other services. When the item DB 407 includes items provided by different services, contents different between different services may be stored in the areas 802 to 808.

  For example, when the item DB 407 includes information on an item provided by the TV service, a value indicating the genre of the TV program is displayed in the area 802, a value indicating the broadcast time zone of the TV program is displayed in the area 803, and a performer of the TV program is displayed in the area 804. May be stored.

  FIG. 9 is an explanatory diagram illustrating the weighting coefficient DB 409 according to the embodiment of this invention.

  The weighting coefficient DB 409 shown in FIG. 9 indicates the weighting coefficient for each attribute group of the user when providing a recipe service.

  The weighting coefficient DB 409 includes areas 901 to 908. Further, the weight coefficient DB 409 illustrated in FIG. 9 includes rows 911 to 917.

  An area 901 and an area 902 indicate groups to which the user belongs. A region 901 illustrated in FIG. 9 includes a group indicating what the user is trying to live in. Further, a region 902 illustrated in FIG. 9 includes a value indicating the user's constitution. The values of the area 901 and the area 902 are specified according to the action history value or the action history value and the personal information value.

  By classifying each user into two groups of an area 901 and an area 902, the server 101 of this embodiment can accurately determine which attribute is important for the user.

  In the weight coefficient DB 409 illustrated in FIG. 9, the user is classified into two group types. However, the server 101 according to the present embodiment may classify the user according to one group type, or more than two. You may classify by the type of many groups.

  Regions 903 to 908 include weighting coefficients corresponding to the regions 803 to 808 of the item DB 407 shown in FIG. Values stored in areas 903 to 908 are stored in advance by the administrator.

  An area 903 shows a weighting factor for menu attributes. An area 904 indicates a weighting factor of the genre attribute. A region 905 shows a weighting factor of the calorie attribute. A region 906 shows the weighting factor of the salinity attribute. A region 907 shows the weighting factor of the protein attribute. An area 908 shows the weighting factor of the cooking time attribute.

  In the present embodiment, an attribute having a high weight coefficient value indicates an attribute that is important for the user.

  Although the weighting factor DB 409 illustrated in FIG. 9 includes the weighting factor assigned to the item attribute, the weighting factor DB 409 of the present embodiment may assign a weighting factor to the attribute value. For example, “1”, “2”, and “3” that are weighting factors may be assigned to the character strings “main meal”, “main dish”, and “dessert” that are attribute values of the “menu”, respectively. .

  The method of setting the weighting coefficient depends on a recommendation calculation method by a recommendation calculation unit 413 described later. Rows 911 to 917 show the weighting factors of the item attributes for the user attribute group.

  FIG. 10 is a flowchart illustrating a process for determining the recommendation level according to the embodiment of this invention.

  The attribute group determination unit 408 starts the process illustrated in FIG. 10 periodically or when requested to recommend an item from the mobile terminal 104. The attribute group determination unit 408 acquires a user who recommends an item from the mobile terminal 104 when starting the process illustrated in FIG. 10.

  Further, the attribute group determination unit 408 may perform the process shown in FIG. 10 for each user indicated by the area 701 of the action history DB 406.

  The attribute group determination unit 408 acquires an action history related to a user who recommends an item from the action history DB 406. Here, the attribute group determination unit 408 acquires one action history from the latest action history in the action history DB 406. Then, 1 is substituted into a counter value i indicating the number of processed action histories (1401).

  When the recommendation device 404 has the personal information DB 405, the attribute group determination unit 408 acquires personal information regarding the user who recommends the item from the personal information DB 405 in step 1401.

  After step 1401, the attribute group determination unit 408 determines whether or not the counter value i is greater than a predetermined action history number (1402).

  Here, the administrator causes the attribute group determination unit 408 to hold in advance the number of action histories necessary for classifying the user into the attribute group as the predetermined action history number in step 1402. Thereby, the attribute group determination unit 408 can acquire a sufficient number of behavior histories for classifying the user into attribute groups.

  If the counter value i is larger than the predetermined action history number, the attribute group determination unit 408 executes Step 1406 because it has acquired a sufficient number of action histories to classify the user into the attribute group.

  When the counter value i is equal to or less than the predetermined number of action histories, the attribute group determination unit 408 executes step 1403 because a sufficient number of action histories for classifying the user into attribute groups has not been acquired.

  The attribute group determination unit 408 determines whether or not the value stored in the action history area 703 acquired in step 1401 indicates a time included in a past period by a predetermined period from the current time.

  Here, the administrator causes the attribute group determining unit 408 to hold in advance a predetermined period of the action history necessary for classifying the user into the attribute group. Thereby, the attribute group determination unit 408 can acquire a sufficiently new action history from the action history DB 406 to classify the user into the attribute group. And the attribute group determination part 408 can determine an attribute group based on a user's newest preference.

  When the value stored in the action history area 703 does not indicate the time included in the past period for a predetermined period from the current time, the attribute group determination unit 408 uses the action history acquired from the action history DB 406 to change the user Are not classified into attribute groups, step 1405 is executed.

  When the value stored in the action history area 703 indicates a time included in the past period for a predetermined period from the current time, the attribute group determination unit 408 determines the value of the action history acquired from the action history DB 406, or Then, the point value of each attribute group is calculated according to the value of the action history and the value of the personal information (1404).

  Here, the attribute group determination unit 408 holds in advance the value of the region 901 or the value of the region 902 corresponding to the values indicated by the region 702, the region 704, the region 705, etc. of the action history. Further, the value of the area 901 or the area 902 corresponding to the values indicated by the areas 605 to 607 of the personal information is stored in advance.

  Then, the attribute group determination unit 408 calculates the point value of the area 901 corresponding to the value of each area of the action history. Further, the attribute group determination unit 408 calculates a point value of the area 902 corresponding to the value of each area of the personal information.

  For example, when recommending an item to the user 1, the attribute group determination unit 408 acquires a row 714 from a row 712 shown in FIG. 7 in step 1401. When the predetermined number of action histories is 3 or more and the value of the area 703 in the row 711 to the row 713 is included within a predetermined period from the current time, the attribute group determination unit 408 determines that the row 711 to the row 711 in step 1404. Using the row 713, the point value of the attribute group is calculated.

  In step 1404, the attribute group determination unit 408 sets the point value of the attribute group “health-oriented” in the area 901 corresponding to “vegetable“ AAA ”” in the area 705 and “purchase” in the area 704 in the row 711. Add the value of. Further, the attribute group determination unit 408 gives a predetermined value to the point value of the attribute group “health orientation” in the area 901 corresponding to “health goods“ BBB ”” in the area 705 in the row 712 and “view” in the area 704. to add. Further, the attribute group determination unit 408 sets a predetermined value for the point value of the attribute group “health orientation” in the area 901 corresponding to “health program“ CCC ”” in the area 705 in the row 713 and “recording” in the area 704. to add.

  Here, the attribute group determination unit 408 holds in advance information indicating that the character strings “vegetable” and “health” included in the region 705 correspond to the attribute group “health-oriented”.

  Further, when the recommendation device 404 has the personal information DB 405 and recommends an item to the user 1, the attribute group determination unit 408 acquires the row 611 of the personal information DB 405 in step 1401. In step 1404, the point value of the area 902 is calculated based on the row 611.

  That is, the attribute group determination unit 408 adds the point value of the attribute group “high blood pressure” in the region 902 corresponding to “high blood pressure” in the region 607 in the row 611 in step 1404.

  Here, the attribute group determination unit 408 holds in advance information indicating that the character string “high blood pressure” included in the region 606 corresponds to the attribute group “high blood pressure”.

  Note that the method of calculating the point value in step 1404 is not limited to the method described above, and any method may be used as long as the user attribute group is determined based on the action history or the action history and personal information.

  After step 1404, the attribute group determination unit 408 adds 1 to the counter value i (1405) and executes step 1402.

  If it is determined in step 1402 that the counter value i is greater than the predetermined number of action histories, the attribute group determination unit 408 selects the attribute group having the highest value among the calculated point values of the attribute group as the user 1. (1406).

  After step 1406, the weighting factor determination unit 410 acquires the weighting factor of the attribute group from the weighting factor DB 409 using the value of the attribute group determined by the attribute group determination unit 408 in step 1406 as a key. Then, the acquired weighting factor is determined as the weighting factor of the recommended item (1407).

  For example, when an item is recommended to the user 1 and the attribute group of the area 901 is determined to be “health-oriented” in step 1406 and the attribute group of the area 902 is determined to be “hypertension” in step 1406, the weight coefficient is determined. In step 1407, the unit 410 acquires the row 913 of the weighting coefficient DB 409. The row 913 shows that the weighting factors of the attributes “menu”, “genre”, “calorie”, “salt”, “protein”, and “cooking time” are “1”, “2”, “7”. , “10”, “5”, and “5”.

  After step 1407, the recommendation calculation unit 413 substitutes 1 for a counter value j indicating the number of items for which the degree of recommendation has been calculated (1408).

  After step 1408, the recommendation calculation unit 413 determines whether or not the counter value j exceeds a predetermined number of recommended item data (1409). Here, the administrator causes the recommendation calculation unit 413 to hold the number of rows in the item DB 407 in advance as the number of recommended item data in step 1409.

  When the counter value j exceeds the predetermined number of recommended item data, the recommendation calculation unit 413 ends the processing shown in FIG. 10 because the recommendation level of all items has been calculated.

  When the counter value j is equal to or less than the predetermined number of recommended item data, the recommendation calculation unit 413 extracts the j-th row of the item DB 407 and normalizes the values of the region 803 to the region 808 of the extracted row (1410). ). Here, the recommendation calculation unit 413 may normalize the values of the region 803 to the region 808 by any method as long as it is converted into a scale comparable to the weighting factor.

  For example, the recommendation calculation unit 413 may replace the character string included in the region 803 with a numerical value corresponding to the character string in order to facilitate the calculation. Further, the recommendation calculation unit 413 may normalize by changing the value included in the region 805 to a lower value as the value included in the region 805 is higher. The recommendation calculation unit 413 may normalize the value by changing the value included in the region 807 to a higher value as the value included in the region 807 is higher.

  The normalization method in step 1410 is predetermined by the administrator.

  After step 1410, the recommendation calculation unit 413 determines the value of the region 803 to the region 808 normalized in the item to be recommended in step 1410 and the value of the region 803 to the region 808 normalized for the purchased or viewed item, for example. The similarity between the recommended item and the purchased or viewed item is calculated (1411). The smaller the distance is, the higher the similarity is, and it is possible to select an item close to the purchased item. When calculating the distance, by reflecting the value of the weighting coefficient determined in step 1407, it is possible to reflect the attribute that the user regards as important. The calculation of the distance reflecting the weight coefficient will be described later.

  After step 1411, the recommendation calculation unit 413 calculates the recommendation level of each item based on the distance calculated in step 1411 (1412). Here, the recommendation calculation unit 413 calculates the reciprocal of the distance as the item recommendation level.

  After step 1412, the recommendation calculation unit 413 adds 1 to the counter value j (1413) and executes step 1409.

  FIG. 11 is an explanatory diagram illustrating the recommendation level according to the embodiment of this invention.

  The value shown in FIG. 11 indicates the degree of recommendation calculated as a result of the processing shown in FIG. The recommendation level calculated by the process shown in FIG. 10 is calculated for each item. The item ID of the area 1201 shown in FIG. 11 corresponds to the value stored in the area 801 of the item DB 407 shown in FIG.

  As the recommendation degree stored in the area 1202 shown in FIG. 8 is larger, the value of the attribute having the higher weight coefficient is closer to the value of the attribute of the purchased or viewed item. That is, an item with a high recommendation degree stored in the area 1202 indicates an item that places importance on attributes required for the attribute group into which the user is classified.

  FIG. 12A is an explanatory diagram illustrating a display screen of recommendation result data 414 according to the embodiment of this invention.

  The recommendation calculation unit 413 generates recommendation result data 414 based on the recommendation degree calculated by the process shown in FIG. Then, the recommendation result data 414 is transmitted to the mobile terminal 104 via the request response processing unit 402 and the request response transmission / reception unit 401.

  Here, the recommendation calculation unit 413 includes identification information (values stored in the area 901 and the area 902 of the 409) in the recommendation result data 414 as a recommendation reason. Further, the recommendation calculation unit 413 includes identification information (a value stored in the area 802 of the item DB 407) indicating a predetermined number of items in descending order of recommendation degree in the recommendation result data 414.

  Further, the recommendation calculation unit 413 may include identification information indicating the attribute of the weighting factor having a larger value among the weighting factors determined in Step 1407 of FIG. 8 in the recommendation result data 414 as the recommendation reason.

  The display unit 504 of the mobile terminal 104 displays a screen 1301 for recommending an item on the display device 305 based on the received recommendation result data 414.

  A screen 1301 in FIG. 12A includes an area 1303 and an area 1302. In the area 1302, the names of items are displayed in descending order of recommendation degree calculated by the process shown in FIG. In the area 1303, the attribute group of the area 901 shown in FIG. 9 is displayed as a recommendation reason.

  By referring to the screen 1301 in FIG. 12A, the user can grasp the recommendation reason indicating that the item is recommended by being classified into what attribute group.

  FIG. 12B is an explanatory diagram illustrating a display screen of recommendation result data 414 according to the embodiment of this invention.

  When the recommendation result data 414 includes identification information indicating the attribute of the weighting coefficient determined in step 1407 of FIG. 8, the display unit 504 may display the screen 1301 of FIG. 12B.

  A screen 1301 in FIG. 12B includes an area 1302 and an area 1304. In the area 1302, the names of items are displayed in descending order of priority.

  In a region 1304, a character string in which a predetermined character string is added to identification information indicating the attribute of the weighting factor included in the recommendation result data 414 is displayed as a recommendation reason. Thereby, the user can grasp the recommendation reason indicating what attribute is regarded as important by being classified into what attribute group.

  The method for calculating the recommendation degree shown in FIG. 10 is not limited to the method for calculating the recommendation degree as long as an attribute having a larger weighting factor can be reflected more greatly by the recommendation degree.

  The recommendation calculation unit 413 may calculate the priority using the evaluation value for the item specified by the user. Thereby, the user can obtain a recommendation result that matches the user's preference.

  FIG. 13 is an explanatory diagram illustrating the item evaluation DB 411 according to the embodiment of this invention.

  The item evaluation DB 411 includes areas 1001 to 1003. It includes the rows 1011 to 1015 shown in FIG.

  The area 1001 includes an identifier that uniquely identifies an item, and corresponds to a value stored in the area 801 of the item DB 407 in FIG. An area 1002 and an area 1003 indicate evaluation values for items evaluated by each user.

  An area 1002 illustrated in FIG. 13 indicates an evaluation value for each item of the user 1. An area 1003 illustrated in FIG. 13 indicates an evaluation value for each item of the user 2. Rows 1011 to 1015 show evaluation values by each user of each item.

  The item evaluation generation unit 412 acquires the user's evaluation value for the item, and stores the evaluation value in the item evaluation DB 411. Specifically, the evaluation value for each item input by the user is stored in the item evaluation DB 411.

  For example, the display unit 504 of the mobile terminal 104 displays a screen for inputting the evaluation value of the item on the display device 305 and transmits the evaluation value input by the user to the server 101. Then, the item evaluation generation unit 412 of the server 101 receives the item evaluation value via the request response transmission / reception unit 401. And the evaluation value of the received item is stored in item evaluation DB411.

  Further, the item evaluation generation unit 412 may store a value in the item evaluation DB 411 based on the action history stored in the action history DB 406. For example, an area 704 and an area 705 in the row 717 of the action history DB 406 illustrated in FIG. 7 indicate that the user 2 has viewed the dish item “FFF”. For this reason, the item evaluation generation unit 412 determines that the user 2 has performed an affirmative operation on the item FFF, and increases the evaluation value for the item FFF of the user 2.

  In step 1412 shown in FIG. 10, the recommendation calculation unit 413 may calculate the item recommendation degree using the distance calculated in step 1411 and the user evaluation value stored in the item evaluation DB 411. This makes it possible to recommend an item that better matches the user's preference.

  For example, in the item evaluation DB 411, when a high evaluation is given to an item with low salinity and the weighting factor of the salinity attribute is high, the recommendation calculation unit 413 uses the evaluation value to determine the recommendation degree of the item with low salinity Can be increased.

  FIG. 14 is an explanatory diagram illustrating a recommendation degree calculation method according to the embodiment of this invention.

  The graph shown in FIG. 14 is a graph in which points indicating each item are arranged according to the attribute value of each item. A point indicated by a cross indicates an item evaluated by the user, and a point indicated by a black circle indicates an item not evaluated by the user.

  The recommendation calculation unit 413 classifies the evaluated items into a high evaluation item group 1101 having a high evaluation and a low evaluation item group 1102 having a low evaluation. Then, an item in the highly evaluated item group 1101 that has the shortest distance from the unrated item 1103 and the item 1104 is selected.

  Then, the recommendation calculation unit 413 determines that the recommendation degree of the item 1104 whose distance from the selected item among the items 1103 and 1104 is larger than the item 1103 is low because the similarity is low, and is selected. The recommendation degree of the item 1103 whose distance from the item is small is determined to be high because the similarity degree is high.

  The recommendation calculation unit 413 calculates the distance between the attribute value of the unevaluated item and the high evaluation item group, and the distance between the attribute value of the unevaluated item and the low evaluation item group, Items whose distance is smaller than the distance from the low evaluation item group may be determined as having a high recommendation level.

  Then, the recommendation calculation unit 413 increases the rank determined as the recommended item by adding a predetermined value to the recommendation degree calculated in step 1412 for the item determined to have a high recommendation degree. May be. That is, the recommendation level may be calculated so that an item determined to have a high recommendation level is easily selected as an item recommended to the user.

  The graph shown in FIG. 14 is a graph illustrated by two attribute values, but the graph of the present embodiment is illustrated by a dimension according to the number of item attributes. In addition, the recommendation calculation unit 413 may use a general Euclidean distance or may use a distance other than the Euclidean distance in order to calculate a distance to another item. Further, a calculation method other than the distance may be used to obtain the similarity between items.

  In addition, after the weighting factor determination unit 410 determines the weighting factor of the item in step 1407, the recommendation calculation unit 413 may multiply or reduce the value of each attribute by multiplying the value of each attribute by the weighting factor. Good. Accordingly, the distance between the items can be calculated by reflecting the weighting factor in the attribute value between the items. Then, it is possible to calculate a recommendation degree reflecting both the evaluation value of the user and the weight coefficient of the item.

  The recommendation calculation unit 413 may calculate the recommendation degree in consideration of not only the evaluation value of the recommendation target user but also the evaluation value of other users. For example, the recommendation level may be calculated by reflecting the evaluation value of the user in the same attribute group as the user recommending the item in the evaluation value of the user recommending the item.

  According to this embodiment, a user attribute group is determined based on a user's behavior history obtained as a result of providing a plurality of services to the user, and the item attribute weighting factor is determined using the determined attribute group as a key. To decide. Then, items recommended to the user are extracted according to the determined weight coefficient. In other words, by assigning an abstract item such as an attribute group to a user and using a weighting factor according to the abstract item, it is possible to determine an optimal item as an item recommended to the user even between different services It becomes.

  Specifically, it is possible to recommend an item of another service to a user who has received only a certain service.

  In addition, according to the present embodiment, which item is regarded as important by the user is acquired based on the behavior history. And it makes it easy to select an item having an attribute close to the attribute of the item that the user attaches importance to as an item to be recommended. This makes it possible to recommend items that match each user's personal preference.

  Moreover, not only items that the user regards as important but also items with attributes similar to the items that the user regards as important can be recommended to the user, and items that match the user's preferences can be recommended .

  Moreover, according to this embodiment, an attribute group is determined based on personal information in addition to the action history. Thereby, the state or preference of each user, which cannot be acquired only by the action history, can be acquired, and the attribute group can be determined. Different items can be recommended depending on the user.

  In addition, according to the present embodiment, when an item determined as a recommended item is presented to the user, information indicating an attribute group into which the user is classified as a recommendation reason is presented to the user. As a result, the user can grasp what kind of group the user is classified into, and can grasp the basis for recommending the item. And the discomfort felt by the user can be reduced.

  Further, according to the present embodiment, information regarding the attribute having the highest weighting coefficient is also presented to the user as the reason for recommendation. Accordingly, the user can grasp which attribute is determined to be important for the user. And this embodiment can reduce the discomfort which a user feels.

  The present invention can be applied to a server for storing and retrieving information, a television, a PC, a recorder, a mobile phone, a mobile terminal, and combinations thereof.

101 server, 102 network, 103 base station, 104 mobile terminal, 201 communication device, 202 processor, 203 storage device, 204 secondary storage device, 301 communication device, 302 processor, 303 input device, 304 storage device, 305 display device 401 Request response transmission / reception unit, 402 Request response processing unit, 403 Recommendation calculation control unit, 404 Recommendation device, 405 Personal information DB, 406 Action history DB, 407 Item DB, 408 Attribute group determination unit, 409 Weight coefficient DB, 410 Weight coefficient determination Unit, 411 item evaluation DB, 412 item evaluation generation unit, 413 recommendation calculation unit, 414 recommendation result data, 501 input unit, 502 request response transmission / reception unit, 503 request response processing , 504 display unit

Claims (10)

An information processing apparatus that receives data from a terminal that provides a service to a user,
The information processing apparatus includes:
With processor, memory and interface,
A transmission / reception unit for storing the acquired action history information in the memory, when acquiring action history information indicating a history of operation of the terminal by the user to receive the service;
Item information including a plurality of candidates for items recommended to the user and attribute values of the candidates,
Weight coefficient information including a weight coefficient indicating the importance of the attribute assigned to a plurality of groups in which the user is classified according to the history;
A group determining unit for determining a group for classifying the user based on the action history information;
A recommendation degree calculation unit that calculates the similarity of attribute values between items based on the weighting factor assigned to the determined group and the item information;
The recommendation degree calculation unit includes:
According to the similarity of attribute values between the items, the recommendation degree of each candidate is calculated,
According to the calculated recommendation level, an item recommended for the user is determined from the candidates. The information processing apparatus includes an evaluation unit that identifies a first item that the user attaches importance to and a second item that the user does not attach importance according to the history indicated by the action history information,
The recommendation degree calculation unit includes:
Candidates whose similarity with the attribute value of the first item is higher than the similarity with the attribute value of the second item based on the weighting factor assigned to the determined group and the item information Is extracted from the item information,
The information processing apparatus according to claim 1, wherein the recommendation degree of the extracted candidate is calculated so that a ranking determined as an item recommended to the user is increased. The information processing apparatus includes:
Having personal information about the user,
The information processing apparatus according to claim 2, wherein the group determination unit determines a group for classifying the user based on the personal information and the action history information. The recommendation degree calculation unit generates recommendation data including identification information indicating an item recommended to the determined user and identification information indicating a group into which the user is classified,
The information processing apparatus according to claim 3, wherein the transmission / reception unit transmits data for displaying identification information included in the generated recommendation data to the terminal. The recommendation degree calculation unit includes:
From the weighting factor information, the weighting factor having the highest importance among the weighting factors assigned to the group into which the user is classified is extracted;
The information processing apparatus according to claim 4, wherein identification information indicating an attribute corresponding to the extracted weight coefficient is included in the recommendation data. An information processing method by an information processing apparatus that receives data from a terminal that provides a service to a user,
The information processing apparatus includes a processor, a memory, and an interface.
The information processing method includes:
When the processor acquires action history information indicating a history of operations of the terminal by the user to receive the service, the acquired action history information is stored in the memory;
The information processing apparatus includes:
Item information including a plurality of candidates for items recommended to the user and attribute values of the candidates,
Weight coefficient information including a weight coefficient indicating the importance of the attribute assigned to a plurality of groups in which the user is classified according to the history, and
The method
The processor determines a group for classifying the user based on the behavior history information;
The processor calculates the similarity of attribute values between items based on the weighting factor assigned to the determined group and the item information;
The processor calculates a recommendation level of each candidate according to the similarity of attribute values between the items,
The information processing method, wherein the processor determines an item to be recommended to the user from the candidates according to the calculated recommendation level. The method
According to the history indicated by the behavior history information, the processor identifies a first item that the user attaches importance to and a second item that the user does not attach importance to,
Based on the weighting factor assigned to the determined group and the item information, the processor has a similarity between the attribute value of the first item and the similarity value with the attribute value of the second item. Higher candidates from the item information,
The information processing method according to claim 6, wherein the processor calculates the recommendation level of the extracted candidates so that a ranking determined as an item recommended to the user is increased. The information processing apparatus has personal information about the user,
The information processing method according to claim 7, wherein the processor determines a group for classifying the user based on the personal information and the action history information. The method
The processor generates recommendation data including identification information indicating an item recommended to the determined user and identification information indicating a group into which the user is classified;
The information processing method according to claim 8, wherein the processor transmits data for displaying identification information included in the generated recommendation data to the terminal. The method
The processor extracts, from the weighting factor information, a weighting factor having the highest importance among weighting factors assigned to the group into which the user is classified,
The information processing method according to claim 9, wherein the processor includes identification information of the attribute corresponding to the extracted weighting factor in the recommendation data.

Images