JP2015225587A - Information provision device, information provision system, information processing terminal, information provision method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an information provision device capable of extracting a variety of information.SOLUTION: An information provision device follows the steps of: extracting relevant information related to items that satisfy a first extraction condition as target user-related information from items related to a target user; extracting relevant information related to items that satisfy a second extraction condition as similar candidate user-related information from items related to users related to the items included in the extracted target user-related information; extracting users similar to the target user from users included in the similar candidate user-related information as similar users based on the target user-related information and the similar candidate user-related information; and extracting relevant information related to items that satisfy a third extraction condition as similar user-related information from items related to the similar users.

Description

  The present invention relates to an information providing apparatus, an information providing system, an information processing terminal, an information providing method, and a program.

  Based on the history information when the target user purchases a product via a website or the like, information indicating the target user's preference is extracted, and the product based on the extracted preference is presented (recommended) to the target user. Equipment is widespread.

  In relation to this, referring to a storage unit that stores use related information in which a use subject identifier, an item identifier, and use time information are associated with each other, the more use related information includes use time information close to a specific time For each item identifier, selection usage related information that is at least part of a plurality of usage related information in the item identifier is selected for each item identifier, and selection usage related information in one item identifier is selected. An information processing apparatus that calculates a degree of association between one item identifier and another item identifier based on selection use related information in another item identifier is known (see Patent Document 1).

JP2013-130921A

  However, the conventional information processing apparatus performs filtering only for either the date / time or the number of selections for both the user or item selection and the selection use related information selection. Different filtering cannot be performed at times, and as a result, it is difficult to perform various filtering.

  Accordingly, the present invention has been made in view of the above-described problems of the prior art, and provides an information providing apparatus, an information providing system, an information processing terminal, an information providing method, and a program capable of performing various information extraction. Is one of the purposes.

  One aspect of the present invention is a target user related information extraction unit that extracts, as target user related information, related information related to an item that satisfies the first extraction condition among items related to a target user, and the extracted target user related information A similar candidate user related information extracting unit that extracts related information related to an item satisfying the second extraction condition among items related to a user related to the item included in the item as similar candidate user related information, the target user related information, and the Based on similar candidate user related information, a similar user extraction unit that extracts a user similar to the target user from among users included in the similar candidate user related information, and among items related to the similar user Related information related to items that satisfy the third extraction condition And similar user related information extraction unit that extracts as Heather-related information is information providing apparatus comprising a.

  ADVANTAGE OF THE INVENTION According to this invention, the information provision apparatus, information provision system, information processing terminal, information provision method, and program which can perform various information extraction can be provided.

It is a figure which shows an example of a structure of the information provision system 1 which concerns on 1st Embodiment. It is a figure which shows an example of the user information table T4 memorize | stored in user information DB4. It is a figure which shows an example of item information table T5 memorize | stored in item information DB5. It is a figure which shows an example of the related information table T6 memorize | stored in related information DB6. It is a figure which shows an example of the recommendation screen P1 displayed on the information processing terminal 3. FIG. 2 is a diagram illustrating an example of a hardware configuration of an information providing apparatus 2. FIG. 3 is a diagram illustrating an example of a functional configuration of an information providing apparatus 2. FIG. 3 is a diagram illustrating an example of a functional configuration of an information processing terminal 3. FIG. It is the figure which represented typically the 1st example of the flow until a recommendation item is extracted by the information provision apparatus 2. FIG. It is the figure which represented typically the 2nd example of the flow until a recommendation item is extracted by the information provision apparatus 2. FIG. It is the figure which represented typically the 3rd example of the flow until a recommendation item is extracted by the information provision apparatus 2. FIG. It is the figure which represented typically the 4th example of the flow until a recommendation item is extracted by the information provision apparatus 2. FIG. It is a flowchart which shows an example of the flow of the process which extracts a recommendation item by the control part. It is a figure which shows an example of the information which shows the presence or absence of the purchase of the item by a user. It is a flowchart which shows an example of the flow of the process which extracts specific relevant information based on extraction conditions. It is a figure which shows an example of the similarity between users, a recommendation index | exponent, and a recommendation item when the conditions for extracting the relevant information after 2013 (an example of the extraction condition concerning date and time) are selected as the first extraction condition. It is a figure which shows an example of the similarity between users, a recommendation index | exponent, and a recommendation item when the extraction conditions for extracting the related information after 2013 (an example of the extraction conditions concerning date and time) are selected as a 2nd extraction condition. . It is a figure which shows an example of the similarity between users, a recommendation index | exponent, and a recommendation item when the relevant information of a specific category is extracted by each of a 1st extraction process, a 2nd extraction process, and a 3rd extraction process. It is a figure which shows an example of a function structure of the information provision apparatus 2a which concerns on 2nd Embodiment. It is a figure which shows an example of the recommendation screen P2 displayed on the information processing terminal 3. FIG. It is the figure which represented typically an example of the flow until a recommendation item is extracted by the information provision apparatus 2a. It is a flowchart which shows an example of the flow of the process which extracts a recommendation item by the control part 26a. It is a figure for demonstrating an example of the process which calculates the similarity between items. It is a flowchart which shows an example of the flow of the process which the control part 26a extracts specific relevant information based on extraction conditions.

<First Embodiment>
Hereinafter, a first embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a diagram illustrating an example of a configuration of an information providing system 1 according to the first embodiment. The information providing system 1 according to the first embodiment includes an information providing device 2, an information processing terminal 3, a user information DB 4, an item information DB 5, and a related information DB 6.

  The information providing system 1 has similar preferences to the target user, for example, based on items that the user of the information processing terminal 3 (hereinafter referred to as the target user) purchases, browses, or reserves via a website or the like. Other users (hereinafter referred to as similar users) are extracted. At the time of this extraction, the information providing system 1 extracts the item based on the purchase history / browsing history / reservation history etc. of the target user recorded by the website etc. (hereinafter collectively referred to as purchase history etc.). Then, similar users are extracted based on the extracted items. Then, the information providing system 1 presents (recommends) to the target user information indicating items that the extracted similar user likes and that are not included in the purchase history of the target user. The information indicating the item presented by the information providing system 1 is, for example, a still image such as a thumbnail image, a moving image, an item name, an explanation about the item, or the like.

  Moreover, the item which the information provision system 1 (or information provision apparatus 2) presents with respect to a target user is called and called a recommendation item. Here, the item is an article such as a book, food, clothing, or electrical appliance, or a service such as cleaning, massage, information provision, delivery, music, movie, animation, TV program, video game, software, web page, etc. Content and Q & A content. In the following, for convenience of explanation, a process for selecting a recommended item for an item purchased in the past based on the purchase history will be described, but a recommended item for an item browsed in the past by the user based on the browsing history can be selected. It is also possible to select a recommended item for an item that the user has left in the past based on the holding history, or a recommended item based on a history that combines some or all of the purchase history, browsing history, and holding history. May be selected.

  The information providing device 2 is, for example, a web server. In response to a request from the information processing terminal 3, the information providing apparatus 2 outputs various display data described in HTML (HyperText Markup Language), CSS (Cascading Style Sheets), and the like to the information processing terminal 3 for display. . Further, the information providing apparatus 2 acquires login information for the target user to log in to a website provided by the information providing apparatus 2 from the information processing terminal 3. The login information is information registered in the user information DB 4 by the target user via the website (that is, via the information providing device 2). For example, the login information includes the user ID of the target user and the user ID. It is information including the associated password and the like.

  When acquiring the login information, the information providing device 2 refers to the user information stored in the user information DB 4 and determines whether or not the user ID (user ID of the target user) included in the login information is registered, If registered, the target user is allowed to log in to the website. The user information is, for example, information in which a user ID is associated with a user name associated with the user ID. The user information may further include information indicating the user's gender, age, residence, information indicating contact information, information indicating preference, and the like. In the following, for convenience of explanation, unless there is a need for distinction, the user indicated by the user ID and the user ID indicating the user are collectively referred to simply as a user.

  When the target user is permitted to log in, the information providing apparatus 2 acquires various information as necessary from the item information DB 5 and the related information DB 6 based on the user ID included in the target user login information. Based on the received information, information indicating recommended items for the target user is extracted. Then, the information providing device 2 generates display data for displaying the extracted recommendation item, and outputs the generated display data to the information processing terminal 3.

  The information processing terminal 3 is, for example, a desktop PC (Personal Computer), a notebook PC, a tablet PC, a mobile phone terminal, a multi-function mobile phone terminal (smartphone), an electronic book reader, a PDA (Personal Digital Assistant), or the like. The information processing terminal 3 acquires the above display data from the information providing device 2 to display various images on the display unit. Further, the information processing terminal 3 outputs login information input from the target user to the website that is the website and displayed on the own apparatus, to the information providing apparatus 2. And the information processing terminal 3 will display the screen based on the acquired display data on an own apparatus, if the display data based on the information which shows a recommendation item from the information provision apparatus 2 are acquired. Below, for convenience of explanation, a screen displayed on the information processing terminal 3 based on display data based on information indicating a recommended item will be referred to as a recommended screen.

  The user information DB 4 is a server that stores the above-described user information. The user information DB 4 outputs user information to the information providing apparatus 2 in response to a request from the information providing apparatus 2. Here, the user information stored in the user information DB 4 will be described with reference to FIG. FIG. 2 is a diagram illustrating an example of the user information table T4 stored in the user information DB4. As shown in FIG. 2, the user information table T4 stores a user ID and a user name in association with each other.

  The item information DB 5 is a server that stores, as item information, information related to items that can be purchased by the target user through the website described above. Item information is information in which an item ID, an item name, a category of the item (for example, a book, an electrical appliance, etc., and may be a novel, a comic, a fan, etc.) are associated with each other. is there. Here, the item information will be described with reference to FIG. FIG. 3 is a diagram illustrating an example of the item information table T5 stored in the item information DB 5. As shown in FIG. 3, the item information table T5 stores an item ID, an item name, a category of the item, and the like in association with each other. In the following, for the sake of convenience of explanation, unless there is a need for distinction, an item indicated by an item ID or an item ID indicating an item may be collectively referred to as an item.

  The related information DB 6 is a server that stores related information related to items purchased by the user in the past through a website or the like where the user ID is registered. The related information is, for example, a purchase history, and is information in which a user ID is associated with information indicating an item purchased by the user, information indicating a date and time when the user purchased the item, and the like. Here, the related information will be described with reference to FIG. FIG. 4 is a diagram illustrating an example of the related information table T6 stored in the related information DB 6. As shown in FIG. 4, the related information table T6 stores a user ID, an item ID for identifying an item, a date and time when the user purchased the item, and the like in association with each other.

  Here, the recommendation screen displayed on the information processing terminal 3 will be described with reference to FIG. FIG. 5 is a diagram illustrating an example of the recommendation screen P <b> 1 displayed on the information processing terminal 3. On the recommendation screen P1, information (for example, thumbnail images and item names in FIG. 5) indicating recommended items for the target user is displayed for each of the categories G1, G2, and G3. For example, a recommendation item related to a plurality of books is displayed in the region R1 where a recommended item related to the category G1 is displayed. Moreover, the recommendation item which concerns on a some comic (manga) is displayed on area | region R2 where the recommendation item which concerns on category G2 is displayed. Moreover, the recommendation item which concerns on several magazines is displayed on area | region R3 where the recommendation item which concerns on category G3 is displayed.

  Note that the recommendation screen P1 shown in FIG. 5 is merely an example. Therefore, the category of the recommendation item displayed on the recommendation screen P1 may be some other category instead of a book, a comic, or a magazine. In FIG. 5, the recommendation items included in each of the regions R1 to R3 are not arranged in the ranking order, but may be displayed in any ranking order. In FIG. 5, thumbnails are displayed as information indicating recommended items. Alternatively, a moving image related to recommended items may be displayed.

  By displaying the recommendation screen P1 as shown in FIG. 5 on the information processing terminal 3, the target user can easily meet his / her preferences simply by logging in to the website without performing a search spontaneously. It is possible to find highly-specific items (that is, recommended items). That is, the information providing system 1 can present recommended items that are highly likely to meet the preferences of the target user. As a result, the information providing system 1 can promote sales of items on a website or the like that sells the recommended items. it can.

  Hereinafter, the hardware configuration of the information providing apparatus 2 will be described with reference to FIG. FIG. 6 is a diagram illustrating an example of a hardware configuration of the information providing apparatus 2. The information providing apparatus 2 includes, for example, a CPU (Central Processing Unit) 21, a storage unit 22, an input receiving unit 23, a display unit 24, and a communication unit 25, and is connected to the network via the communication unit 25. Communicates with various servers and devices. These components are connected to each other via a bus Bus so that they can communicate with each other. The CPU 21 executes various programs stored in the storage unit 22.

  The storage unit 22 includes, for example, an HDD (Hard Disk Drive), an SSD (Solid State Drive), an EEPROM (Electrically Erasable Programmable Read-Only Memory), a ROM (Read-Only Memory), a RAM (Random Access Memory), and the like. Various information, images, and programs processed by the information providing apparatus 2 are stored. The storage unit 22 may be an external storage device connected via a digital input / output port such as a USB instead of the one provided in the information providing device 2.

The input receiving unit 23 is, for example, a keyboard, mouse, touch pad, or other input device. The input receiving unit 23 may be configured as a touch panel together with the display unit 24.
The display unit 24 is, for example, a liquid crystal display panel or an organic EL (ElectroLuminescence) display panel.
The communication unit 25 includes, for example, a digital input / output port such as USB, an Ethernet (registered trademark) port, and the like.

  The information processing terminal 3 has a hardware configuration similar to the hardware configuration shown in FIG. 6. In the following, in order to distinguish between the hardware configuration of the information providing device 2 and the hardware configuration of the information processing terminal 3, These are referred to as CPU 31, storage unit 32, input reception unit 33, display unit 34, and communication unit 35.

  Hereinafter, the functional configuration of the information providing apparatus 2 will be described with reference to FIG. FIG. 7 is a diagram illustrating an example of a functional configuration of the information providing apparatus 2. The information providing apparatus 2 includes a storage unit 22, an input receiving unit 23, a communication unit 25, and a control unit 26. Some or all of these functional units are realized by, for example, the CPU 21 executing various programs stored in the storage unit 22. Some or all of these functional units may be hardware functional units such as LSI (Large Scale Integration) and ASIC (Application Specific Integrated Circuit).

  The control unit 26 includes a user related information extraction condition determining unit 261, a target user related information extracting unit 262, a similar candidate user extracting unit 263, a similar candidate user related information extracting unit 264, and an inter-user similarity calculating unit 265. , A similar user extraction unit 266, a similar user related information extraction unit 267, a recommendation index calculation unit 268, a recommendation item extraction unit 269, and a recommendation item output unit 270.

  In response to a request from the information processing terminal 3, the control unit 26 outputs various types of information related to the website provided by the information providing device 2 to the information processing terminal 3 via the communication unit 25. Further, when the control unit 26 acquires the login information from the information processing terminal 3, the control unit 26 refers to the user information stored in the user information DB 4 via the communication unit 25, and the user ID of the target user included in the login information is registered. If it is registered, the target user is allowed to log in to the website. The control unit 26 is an example of an acquisition unit.

  The user related information extraction condition determination unit 261 is stored in the storage unit 22 in response to requests from the target user related information extraction unit 262, the similar candidate user related information extraction unit 264, and the similar user related information extraction unit 267, respectively. Load multiple user-related information extraction conditions. The user related information extraction condition is related information stored in the related information DB 6, and is related to each of the target user related information extraction unit 262, similar candidate user related information extraction unit 264, and similar user related information extraction unit 267. This is a condition for extracting desired related information (more specifically, related information representing the user's preference) from the information.

  For example, a plurality of user related information extraction conditions include an extraction condition related to date and time (for example, a condition for extracting related information related to an item purchased in the past year), an extraction condition related to a category (for example, a book Conditions for extracting related information), extraction conditions related to the number of extractions (for example, conditions for extracting only the latest 10 items), conditions for extracting all (no specific related information is extracted) Condition) and the like.

  The item included in the related information extracted by the extraction condition related to the date and time reflects the user's preference in a specific period indicated by the extraction condition. The item included in the related information extracted by the extraction condition related to the category reflects the user's preference in the specific category indicated by the extraction condition. The items included in the related information extracted by the extraction condition relating to the number of extractions reflect the user's preference during the period (for example, the most recent week) when those items were purchased. In addition, a specific preference is not reflected in the item included in the related information extracted by the conditions for extracting all. Hereinafter, for convenience of explanation, the conditions for extracting all are referred to as all extraction conditions, and other user-related information extraction conditions other than all extraction conditions (for example, the above-described extraction conditions related to date and time, extraction conditions related to categories, etc. ) Will be referred to as a specific extraction condition unless there is a need for distinction.

  The user related information extraction condition determination unit 261 selects the read user related information extraction condition according to a predetermined selection rule. The predetermined selection rule is, for example, selecting at random from a plurality of user-related information extraction conditions or selecting in a predetermined order. Further, when selecting the user related information extraction condition, the user related information extraction condition determination unit 261 selects a different number of user related information extraction conditions from the plurality of user related information extraction conditions each time the user related information extraction condition is selected.

  Note that the user-related information extraction condition determination unit 261 always extracts the same number of user-related information instead of a configuration in which a different number of user-related information extraction conditions are selected from the plurality of user-related information extraction conditions each time selection is made. A configuration for selecting conditions may be used, or a configuration for periodically changing the number to be selected may be used. The user-related information extraction condition determination unit 261 outputs the selected user-related information extraction condition to the requesting function unit. The user related information extraction condition determination unit 261 is an example of an extraction condition determination unit.

  The target user related information extraction unit 262 acquires related information associated with the user ID of the target user from the related information DB 6 based on the user ID of the target user. Then, the target user related information extraction unit 262 causes the user related information extraction condition determination unit 261 to select a user related information extraction condition. Hereinafter, for convenience of explanation, the user related information extraction condition selected by the user related information extraction condition determination unit 261 in response to a request from the target user related information extraction unit 262 will be referred to as a first extraction condition.

  The target user related information extraction unit 262 acquires the first extraction condition selected by the user related information extraction condition determination unit 261, and acquires the first extraction condition acquired from the related information associated with the user ID of the target user. Related information that satisfies the conditions is extracted as target user related information. Hereinafter, for convenience of explanation, the process of extracting the target user related information will be referred to as a first extraction process. For example, when the first extraction condition is a condition for extracting related information related to a book, the target user related information extracting unit 262 extracts related information related to the book purchased by the target user as first related information.

  In addition, the target user related information extraction unit 262 is, for example, a condition for the first extraction condition to extract related information related to a book and a condition for extracting related information related to an item purchased in the past one year. In this case, related information relating to books purchased by the target user in the past year is extracted as target user related information. As described above, the target user related information extraction unit 262 can extract the target user related information according to the first extraction condition selected by the user related information extraction condition determination unit 261. As a result, the target user related information extraction unit 262 can extract different target user related information each time its own function is executed. The target user related information extraction unit 262 outputs the extracted target user related information to the similar candidate user extraction unit 263 and the inter-user similarity calculation unit 265.

  When the similar candidate user extraction unit 263 acquires the target user related information from the target user related information extraction unit 262, the similar candidate user extraction unit 263 refers to the acquired target user related information and purchases an item included in the target user related information in the past. For a certain user, the user ID of the user who purchased the item indicated by the item ID included in the target user related information is read (extracted) from the related information DB 6. By this extraction, the similar candidate user extraction unit 263 can extract a user who is highly likely to have a similar preference to the target user as a similar candidate user. The similar candidate user extraction unit 263 outputs a user ID indicating the extracted similar candidate user to the similar candidate user related information extraction unit 264.

  When acquiring the user ID indicating the similar candidate user, the similar candidate user related information extraction unit 264 acquires the related information associated with the user ID from the related information DB 6 based on the acquired user ID. Then, the similar candidate user related information extraction unit 264 causes the user related information extraction condition determination unit 261 to select a user related information extraction condition. Hereinafter, for convenience of explanation, the user related information extraction condition selected by the user related information extraction condition determination unit 261 in response to a request from the similar candidate user related information extraction unit 264 will be referred to as a second extraction condition.

  The similar candidate user related information extraction unit 264 acquires the second extraction condition selected by the user related information extraction condition determination unit 261, and acquires the second extraction condition acquired from the related information including the user ID indicating the similar candidate user. Related information that satisfies the above is extracted as similar candidate user related information. Hereinafter, for convenience of explanation, the process of extracting similar candidate user related information will be referred to as a second extraction process. For example, when the second extraction condition is a condition for extracting related information related to a book, the similar candidate user related information extracting unit 264 uses related information related to the book purchased by the similar candidate user as similar candidate user related information. Extract.

  In addition, the similar candidate user related information extraction unit 264 is, for example, a condition for the second extraction condition to extract related information related to a book and a condition for extracting related information related to an item purchased in the past one year. In this case, the related information related to the book purchased by the similar candidate user in the past year is extracted as the similar candidate user related information. As described above, the similar candidate user related information extraction unit 264 can extract the similar candidate user related information corresponding to each second extraction condition selected by the user related information extraction condition determination unit 261. As a result, the similar candidate user related information extraction unit 264 can extract different similar candidate user related information each time its own function is executed. The similar candidate user related information extraction unit 264 outputs the extracted similar candidate user related information to the inter-user similarity calculation unit 265.

  The inter-user similarity calculation unit 265 acquires the target user related information from the target user related information extraction unit 262. Further, the inter-user similarity calculation unit 265 acquires similar candidate user related information from the similar candidate user related information extraction unit 264. Based on the acquired target user related information and similar candidate user related information, the inter-user similarity calculation unit 265 calculates an inter-user similarity that represents the similarity with the target user for each similar candidate user. The inter-user similarity calculation unit 265 outputs the calculated inter-user similarity for each similar candidate user to the similar user extraction unit 266.

  The similar user extraction unit 266 acquires the similarity between users for each similar candidate user from the similarity calculation unit 265 between users. The similar user extraction unit 266 extracts, as similar users, users who have similar preferences to the target user from the similar candidate users according to a predetermined rule based on the acquired similarity between users for each similar candidate user. The predetermined rules include, for example, extracting similar candidate users as similar users in descending order of similarity between users, extracting similar candidate users having a similarity between users equal to or higher than a predetermined threshold as similar users, etc. It is. The similar user extraction unit 266 outputs the extracted user ID of the similar user to the similar user related information extraction unit 267.

  When the similar user related information extracting unit 267 acquires the user ID of the similar user from the similar user extracting unit 266, the similar user related information extracting unit 267 reads the related information associated with the user ID from the related information DB 6 based on the acquired user ID. Then, the similar user related information extraction unit 267 causes the user related information extraction condition determination unit 261 to select a user related information extraction condition. Hereinafter, for convenience of explanation, the user related information extraction condition selected by the user related information extraction condition determination unit 261 in response to a request from the similar user related information extraction unit 267 will be referred to as a third extraction condition.

  The similar user related information extraction unit 267 acquires the third extraction condition selected by the user related information extraction condition determination unit 261, and acquires the third extraction condition acquired from the related information including the acquired user ID of the similar user. The related information satisfying the condition is extracted as similar user related information. Hereinafter, for convenience of explanation, the process of extracting similar user related information will be referred to as a third extraction process. For example, when the third extraction condition is a condition for extracting related information related to a book, the similar user related information extracting unit 267 extracts related information related to a book purchased by a similar user as similar user related information.

  In addition, the similar user related information extraction unit 267 is, for example, a condition for the third extraction condition to extract related information related to a book and a condition for extracting related information related to an item purchased in the past one year. In this case, the related information related to the book purchased by the target user in the past year is extracted as similar user related information. As described above, the similar user related information extraction unit 267 can extract similar user related information according to the third extraction condition selected by the user related information extraction condition determination unit 261. As a result, the similar user related information extraction unit 267 can extract different similar user related information every time its own function is executed. The similar user related information extraction unit 267 outputs the extracted similar user related information and the similarity between users for each similar user to the recommendation index calculation unit 268.

  The recommendation index calculation unit 268 acquires the similar user related information and the similarity between users for each similar user from the similar user related information extraction unit 267. The recommendation index calculation unit 268 calculates a recommendation index for each item included in the similar user related information based on the acquired similar user related information and the similarity between users for each similar user. The recommendation index is a value indicating importance when providing information to a target user. An item with a large recommendation index value can be highly promoted by providing information to the target user. The recommended index calculating unit 268 outputs the calculated recommended index to the recommended item extracting unit 269 together with similar user related information.

  The recommended item extraction unit 269 acquires a recommendation index for each item included in the similar user related information and the similar user related information from the recommended index calculation unit 268. The recommended item extraction unit 269 extracts an item that becomes a recommended item from the third related information according to a predetermined rule based on the acquired recommended index and the recommended index for each item included in the similar user related information.

  The predetermined rule is, for example, extracting a predetermined number or items (for example, about 10 items) input by the user in the order of items with a high recommendation index, or a recommendation index equal to or higher than a predetermined threshold. For example, extracting a predetermined number of items at random, or the like. The recommended item extraction unit 269 outputs the item ID of the extracted recommended item to the recommended item output unit 270 together with the recommendation index for each item ID.

  The recommended item output unit 270 acquires the item ID of the recommended item from the recommended item extraction unit 269 together with the recommendation index for each item ID. And the recommendation item output part 270 acquires the information (for example, a thumbnail image, a moving image, etc.) which concerns on the recommendation item which acquired item ID shows from item information DB5, and acquired information in order based on the recommendation index. By arranging, information representing a recommendation screen as shown in FIG. 5 is generated.

  The recommended item output unit 270 outputs information representing the generated recommendation screen to the information processing terminal 3 via the communication unit 25. The order based on the recommendation index is the order in which the recommendation index is high or low. Further, the recommended item output unit 270 may be configured to generate a recommendation screen by arranging information acquired from the item information DB 5 in the order selected at random. The recommended item output unit 270 is an example of a first output unit.

  Hereinafter, the functional configuration of the information processing terminal 3 will be described with reference to FIG. FIG. 8 is a diagram illustrating an example of a functional configuration of the information processing terminal 3. The information processing terminal 3 includes an input reception unit 33, a display unit 34, a communication unit 35, and a control unit 36. Some or all of these functional units are realized, for example, by the CPU 31 executing various programs stored in the storage unit 32. Some or all of these functional units may be hardware functional units such as LSI (Large Scale Integration) and ASIC (Application Specific Integrated Circuit).

The control unit 36 includes a display control unit 37. The control unit 36 receives input from the user by the input receiving unit 33 and outputs information indicated by the received input to the information providing apparatus 2 via the communication unit 35. The control unit 36 is an example of a second output unit.
The display control unit 37 causes the display unit 34 to display a display image based on the display data acquired from the information providing apparatus 2 via the communication unit 35.

  Hereinafter, with reference to FIGS. 9 to 12, a flow until a recommended item is extracted by the information providing apparatus 2 will be described. FIG. 9 is a diagram schematically illustrating a first example of a flow until a recommendation item is extracted by the information providing apparatus 2. In the following, for convenience of explanation, all the processes by the functional units of the control unit 26 will be described as the processes of the control unit 26, but actually the processes performed by the functional units described in FIG. The example shown in FIG. 9 is an example in which the first extraction condition is a specific extraction condition, and the second extraction condition and the third extraction condition are all extraction conditions.

  First, the control unit 26 performs the first extraction process based on the first extraction condition based on the user ID of the target user U1, so that the target user U1 purchases the item that satisfies the first extraction condition in the past. The item (currently owned) is extracted as the target item I1 (step S1). Here, the first extraction condition in the example illustrated in FIG. 9 is a condition for extracting items purchased in the past year. That is, an item that satisfies the first extraction condition is an item that reflects the preference of the target user U1 for the past year (in other words, a partial preference among all the preferences of the target user). Based on the first extraction condition, the control unit 26 extracts the item purchased by the target user U1 in the past year as the target item I1.

  Next, the control part 26 extracts the user who purchased the target item I1 extracted at step S1 in the past as the similar candidate user U2 (step S2). Next, the control part 26 performs the 2nd extraction process based on 2nd extraction conditions, and the item which satisfy | fills 2nd extraction conditions and the similar candidate user U2 purchased in the past as the similar candidate item I2 Extract (step S3).

  Here, it is assumed that the second extraction condition in the example shown in FIG. That is, the items satisfying the second extraction condition are all items purchased by the similar candidate user U2 in the past. Based on the second extraction condition, the control unit 26 extracts all items previously purchased by the similar candidate user U2 as similar candidate items I2. Here, the similar candidate item I2 extracted in step S3 is an item reflecting the preference of the similar candidate user U2.

  Next, the control unit 26 calculates the similarity between users based on the target item I1 extracted in step S1 and the similar candidate item I2 extracted in step S3 (step S4). Here, the similarity between users is calculated based on the degree of association between an item purchased by the target user in the past and an item purchased by the similar candidate user in the past, so the preference of the target user and the similar candidate user It is a value that reflects both tastes. Next, the control unit 26 extracts a similar user U3 from the similar candidate users U2 based on the similarity between users calculated in step S4 (step S5).

  Next, the control unit 26 performs the third extraction process based on the third extraction condition, thereby extracting an item satisfying the third extraction condition and purchased by the similar user U3 in the past as the similar item I3. (Step S6). Here, it is assumed that the third extraction condition in the example illustrated in FIG. That is, the items satisfying the third extraction condition are all items purchased by the similar user U3 in the past. That is, the control unit 26 extracts all items purchased by the similar user U3 in the past as the similar items I3 based on the third extraction condition. Here, the similar item I3 extracted in step S6 is an item reflecting the preference of the similar user U3.

  Next, the control unit 26 calculates a recommendation index of the similar item I3 based on the similarity between users calculated in step S4 and the similar item I3 extracted in step S6 (step S7). The control unit 26 selects the recommended item I4 based on the recommendation index calculated in step S7 (step S8). Here, the recommendation index is, for example, a sum of similarities between similar users with respect to items purchased by each similar user in the past. For this reason, the recommendation index reflects all of the preferences of the target user, the preferences of the similar candidate users, and the preferences of the similar users. Therefore, the item I4 selected based on the recommendation index includes all of the target user's preference (in this example, the target user's preference over the past year), the similar candidate user's preference, and the similar user's preference. It is reflected.

  As described above, the control unit 26 selects all the extraction conditions as the second extraction condition and the third extraction condition, and selects the specific extraction condition as the first extraction process, thereby resembling the partial preference of the target user. The recommendation item I4 reflecting the preference of the candidate user and the similar user can be selected.

  FIG. 10 is a diagram schematically illustrating a second example of a flow until a recommended item is extracted by the information providing apparatus 2. The example illustrated in FIG. 10 is an example in which the second extraction condition is a specific extraction condition and the first extraction condition and the third extraction condition are all extraction conditions. First, the control unit 26 performs the first extraction process based on the first extraction condition based on the user ID of the target user U4, so that the target user U4 purchases the item that satisfies the first extraction condition in the past. The item (currently owned) is extracted as the target item I5 (step S11).

  Here, it is assumed that the first extraction condition in the example shown in FIG. That is, the items satisfying the first extraction condition are all items purchased by the target user U4 in the past. Based on the first extraction condition, the control unit 26 extracts all items purchased in the past by the target user U4 as the target item I5. Therefore, the target item I5 extracted in step S11 is an item reflecting the preference of the target user U4.

  Next, the control part 26 extracts the user who purchased the target item I5 extracted at step S11 in the past as the similar candidate user U5 (step S12). Next, the control part 26 extracts the item which the similar candidate user U5 purchased in the past as the similar candidate item I6 by performing the 2nd extraction process based on 2nd extraction conditions (step S13).

  Here, the second extraction condition in the example shown in FIG. 10 is a condition for extracting items purchased in the past year. That is, the item that satisfies the second extraction condition is an item that reflects the preference of the similar candidate user U5 over the past year. Based on the second extraction condition, the control unit 26 extracts the item purchased by the similar candidate user U5 in the past year as the similar candidate item I6. Accordingly, the similar candidate item I6 extracted in step S13 is an item reflecting the preference of the similar candidate user U5 for the past year (in other words, a partial preference among all the preferences of the similar candidate user U5). .

  Next, the control unit 26 calculates the similarity between users based on the target item I5 extracted in step S11 and the similar candidate item I6 extracted in step S13 (step S14). Next, the control unit 26 extracts a similar user U6 from the similar candidate users U5 based on the similarity between users calculated in step S14 (step S15). Next, the control part 26 performs the 3rd extraction process based on 3rd extraction conditions, and extracts the item which satisfy | fills 3rd extraction conditions and the similar user U6 purchased in the past as the similar item I7. (Step S16).

  Here, it is assumed that the third extraction condition in the example illustrated in FIG. That is, the items satisfying the third extraction condition are all items purchased by the similar user U6 in the past. Based on the third extraction condition, the control unit 26 extracts all items purchased by the similar user U6 in the past as similar items I7. Therefore, the similar item I7 extracted in step S16 is an item reflecting only the preference of the similar user U6.

  Next, the control unit 26 calculates a recommendation index of the similar item I7 based on the similarity between users calculated in step S14 and the similar item I7 extracted in step S16 (step S17). The control unit 26 selects the recommended item I8 based on the recommendation index calculated in step S17 (step S18).

  Here, since the recommendation item I8 selected in step S18 is selected based on the recommendation index, the preference of the target user and the preference of the similar candidate user (in this example, the past one year of the similar candidate user). (Preferences) and preferences of similar users are all reflected. In this way, the control unit 26 selects the entire extraction condition as the first extraction condition and the third extraction condition, and selects the specific extraction condition as the second extraction process, whereby the partial preference of the similar candidate user, The recommended item I8 reflecting the preferences of the target user and similar users can be selected.

  FIG. 11 is a diagram schematically illustrating a third example of a flow until a recommended item is extracted by the information providing apparatus 2. In addition, the example illustrated in FIG. 11 is an example in which the first extraction condition and the third extraction condition are specific extraction conditions, and the second extraction condition is all extraction conditions. First, the control unit 26 performs the first extraction process based on the first extraction condition based on the user ID of the target user U7, so that the target user U7 purchases the item that satisfies the first extraction condition in the past. The item currently owned (currently owned) is extracted as the target item I9 (step S1).

  Here, it is assumed that the first extraction condition in the example shown in FIG. 11 is a condition for extracting items purchased in the past year. That is, an item that satisfies the first extraction condition is an item that reflects the preferences of the target user U7 over the past year. Based on the first extraction condition, the control unit 26 extracts the item purchased by the target user U7 in the past year as the target item I9.

  Next, the control unit 26 extracts a user who has purchased the target item I9 extracted in step S21 in the past as the similar candidate user U8 (step S22). Next, the control part 26 extracts the item which the similar candidate user U8 purchased in the past as the similar candidate item I10 by performing the 2nd extraction process based on 2nd extraction conditions (step S23).

  Here, it is assumed that the second extraction condition in the example illustrated in FIG. That is, the items satisfying the second extraction condition are all items purchased by the similar candidate user U8 in the past. Based on the second extraction condition, the control unit 26 extracts all items purchased by the similar candidate user U8 in the past as similar candidate items I10. Therefore, the similar candidate item I2 extracted in step S23 is an item reflecting the preference of the similar candidate user U8.

  Next, the control unit 26 calculates the similarity between users based on the target item I9 extracted in step S21 and the similar candidate item I10 extracted in step S23 (step S24). Next, the control unit 26 extracts a similar user U9 from the similar candidate users U8 based on the similarity between users calculated in step S24 (step S25).

  Next, the control part 26 performs the 3rd extraction process based on 3rd extraction conditions, The item which satisfy | fills 3rd extraction conditions among the items which the similar user U9 extracted by step S25 purchased in the past is shown. Extract (step S26). Here, the third extraction condition in the example illustrated in FIG. 11 is a condition for extracting items purchased in the past year. That is, the item satisfying the third extraction condition is an item reflecting the preference of the similar user U9 for the past year. Based on the third extraction condition, the control unit 26 extracts the item purchased by the similar user U9 in the past year as the similar item I11.

  Next, the control unit 26 calculates a recommendation index of the similar item I11 based on the similarity between users calculated in step S24 and the similar item I11 extracted in step S25 (step S27). The control unit 26 selects the recommended item I12 based on the recommendation index calculated in step S27 (step S28).

  Here, since the recommended item I12 selected in step S28 is selected based on the recommendation index, the preference of the target user (in this example, the preference of the target user for the past year) and the similar candidate user's preference All of the preferences and preferences of similar users (in this example, preferences of similar users over the past year) are reflected. As described above, the control unit 26 selects all the extraction conditions as the second extraction conditions, and selects the specific extraction conditions as the first extraction condition and the third extraction process, thereby similar to the partial preference of the target user. It is possible to select a recommendation item that reflects the partial preference of the user and the preference of the similar candidate user.

  FIG. 12 is a diagram schematically illustrating a fourth example of a flow until a recommendation item is extracted by the information providing apparatus 2. The example illustrated in FIG. 12 is an example in which the second extraction condition and the third extraction condition are specific extraction conditions, and the first extraction condition is all extraction conditions. First, the control unit 26 performs the first extraction process based on the first extraction condition based on the user ID of the target user U10, so that the target user U10 purchases the item that satisfies the first extraction condition in the past. The item currently owned (existing) is extracted as the target item I13 (step S31).

  Here, it is assumed that the first extraction condition in the example shown in FIG. That is, the items that satisfy the first extraction condition are all items that the target user U10 has purchased in the past. The control unit 26 extracts all items purchased by the target user U10 in the past as the target item I13 based on the first extraction condition. Accordingly, the target item I13 extracted in step S31 is an item reflecting the target user's preference.

  Next, the control part 26 extracts the user who purchased the target item I13 extracted at step S31 in the past as the similar candidate user U11 (step S32). Next, the control part 26 extracts the item which the similar candidate user U11 purchased in the past as the similar candidate item I14 by performing the 2nd extraction process based on 2nd extraction conditions (step S33).

  Here, the second extraction condition in the example shown in FIG. 12 is a condition for extracting items purchased in the past year. That is, the item satisfying the second extraction condition is an item reflecting the preference of the similar candidate user U11 for the past year. Based on the second extraction condition, the control unit 26 extracts the item purchased by the similar candidate user U11 in the past year as the similar candidate item I14.

  Next, the control unit 26 calculates the similarity between users based on the target item I13 extracted in step S31 and the similar candidate item I14 extracted in step S33 (step S34). Next, the control unit 26 extracts a similar user U12 from the similar candidate users U11 based on the similarity between users calculated in step S34 (step S35). Next, the control unit 26 performs the third extraction process based on the third extraction condition to extract an item satisfying the third extraction condition and purchased in the past by the similar user U12 as the similar item I15. (Step S36).

  Here, the third extraction condition in the example shown in FIG. 12 is a condition for extracting items purchased in the past year. That is, an item that satisfies the third extraction condition is an item that reflects the preferences of the similar user U12 over the past year. Based on the third extraction condition, the control unit 26 extracts the item purchased by the similar user U12 in the past year as the similar item I15.

  Next, the control unit 26 calculates a recommendation index of the similar item I15 based on the similarity between users calculated in step S34 and the similar item I15 extracted in step S35 (step S37). The control unit 26 selects a recommendation item I16 based on the recommendation index calculated in step S37 (step S38).

  Here, since the recommended item I16 selected in step S38 is selected based on the recommendation index, the preference of the target user and the preference of the similar candidate user (in this example, the preference of the target user for the past year). ) And preferences of similar users (in this example, preferences of similar users over the past year) are reflected. In this way, the control unit 26 selects the entire extraction condition as the first extraction condition, and selects the specific extraction condition as the second extraction condition and the third extraction process, whereby the preference of the target user and the similar user part The recommended item I16 reflecting the general preference and the partial preference of the similar candidate user can be selected.

  As described with reference to FIGS. 9 to 12, the control unit 26 selects the specific extraction condition and the total extraction condition for each of the first extraction process to the third extraction process every time a recommended item is selected. A recommendation item reflecting a different preference can be selected by randomly combining a part or all of. As a result, the control unit 26 can provide (present) a different recommended item each time for the target user, and can obtain a high sales promotion effect. In particular, as in the example illustrated in FIG. 12, the control unit 26 can present a recommended item selected based on a preference different from the preference of the target user to the target user. As a result, the target user can discover his / her preferences that he / she does not even notice.

  In addition, with respect to the first extraction process to the third extraction process in the control unit 26, some or all combinations of the specific extraction conditions and the total extraction conditions are not limited to the four patterns illustrated in FIGS. Any other combination may be used. In addition, each of the first extraction process to the third extraction process may be performed based on a plurality of extraction conditions. For example, in the example illustrated in FIG. 9, for the first extraction condition in the first extraction process In addition to the extraction condition related to the date and time, the extraction condition related to the category may be selected.

  Hereinafter, with reference to FIG. 13, the process which extracts the recommendation item by the control part 26 is demonstrated. FIG. 13 is a flowchart illustrating an example of a process flow for extracting recommended items by the control unit 26. In the following description, it is assumed that the control unit 26 has already acquired login information from the information processing terminal 3 and has permitted the target user to log in. First, the target user related information extraction unit 262 acquires related information including the user ID of the target user from the related information DB 6 based on the user ID of the target user. Then, the target user related information extraction unit 262 causes the user related information extraction condition determination unit 261 to select the first extraction condition, and selects the selected first extraction condition from the related information including the acquired user ID of the target user. The related information to be satisfied is extracted as target user related information (that is, items previously purchased by the target user are extracted) (step S100).

  Next, the similar candidate user extraction unit 263 sets a user who has purchased an item indicated by the item ID included in the target user related information extracted by the target user related information extraction unit 262 in the past as a similar candidate user. User IDs of candidate users are extracted (step S110). Next, the similar candidate user related information extraction unit 264 acquires related information including the user ID of the similar candidate user from the related information DB 6 based on the user ID of the similar candidate user. Then, the similar candidate user related information extraction unit 264 causes the user related information extraction condition determination unit 261 to select the second extraction condition, and selects the second extraction selected from the related information including the acquired user ID of the similar candidate user. Related information that satisfies the conditions is extracted as similar candidate user related information (that is, items previously purchased by the similar candidate user are extracted) (step S120).

  Next, the inter-user similarity calculation unit 265 is based on the target user related information extracted by the target user related information extraction unit 262 and the similar candidate user related information extracted by the similar candidate user related information extraction unit 264. The similarity between users is calculated (step S130). Here, processing for calculating the similarity between users will be described. When the set of items purchased by the user K is X and the set of items purchased by the user L is Y, the inter-user similarity calculation unit 265 sets the Jaccard coefficient based on the following equation (1) as the inter-user similarity sim. calculate.

sim = | X∩Y | / | X∪Y | (1)

  The Jaccard coefficient can be used when the information included in the related information and indicating whether or not an item has been purchased by the user is expressed as 1 for purchase and 0 for no purchase. Note that the inter-user similarity calculation unit 265 may be configured to calculate some other coefficient as the inter-user similarity instead of the Jaccard coefficient when calculating the inter-user similarity. For example, the inter-user similarity calculation unit 265 may be configured to calculate a Cosine coefficient based on the following equation (2) as the inter-user similarity sim.

sim = | X∩Y | / sqrt (| X || Y |) (2)

  In addition, information included in the related information is expressed as information indicating an evaluation of an item by the user (for example, 0 for unpurchased items, five-level evaluation of 1 to 5 for satisfaction). In this case, the inter-user similarity calculation unit 265 may be configured to calculate the Pearson correlation coefficient, the Cosine coefficient shown in the above equation (2), and the like as the inter-user similarity sim.

  Here, with reference to FIG. 14, the similarity between users calculated by the similarity calculation part 265 between users is demonstrated. FIG. 14 is a diagram illustrating an example of information indicating whether or not an item has been purchased by a user. The table T-ui1 is set so that each of the six users indicated by the user IDs u1 to u6 purchases each of the items indicated by the item IDs i1 to i6, and is left blank when not purchased. Represents information indicating whether or not an item has been purchased. In FIG. 14, it is assumed that the user indicated by the user ID ui1 is the target user.

  Based on the target user related information and the similar candidate user related information, the inter-user similarity calculation unit 265 extracts and extracts information indicating whether or not an item has been purchased by the user as represented by the table T-ui1. Based on the information indicating whether or not an item has been purchased by the user, the Jaccard coefficient is calculated as the similarity between users by the above equation (1). In the column C1 in FIG. 14, based on information indicating whether or not the user has purchased an item represented by the table T-ui1, the similarity calculation unit 265 between users is similar candidate users (in this example, user IDs u2 to u6). The result of calculating the similarity between the target user and the target user is shown.

  As shown in the column C1, the inter-user similarity sim (u2) to sim (u4) between the target user and each of the similar candidate users indicated by the user IDs u2 to u4 is 0.50. Also, the inter-user similarity sim (u5) to sim (u6) between the target user and each of the similar candidate users indicated by the user IDs u5 to u6 is 0.20, respectively.

  Next, the similar user extraction unit 266 extracts a similar user from the similar candidate users based on the inter-user similarity calculated by the inter-user similarity calculation unit 265 in step S130 (step S140). Here, processing for extracting similar users from similar candidate users by the similar user extracting unit 266 will be described. The similar user extraction unit 266 extracts a user ID of a similar user by setting a predetermined number of similar candidate users as similar users in descending order of the degree of similarity between users calculated in step S130. In this example, the predetermined number is 10. However, the predetermined number is not limited to this number, and may be any number as long as it is one or more. Therefore, in the example illustrated in FIG. 14, the similar user extraction unit 266 extracts all five similar candidate users indicated by the user IDs u2 to u6 as similar users.

  Instead of this, the similar user extraction unit 266 may be configured such that, for example, similar candidate users whose similarity between users is higher than a predetermined value are similar users, and all similar candidate users are similar users. A configuration may be used, and a configuration may be adopted in which some or all of the similar candidate users are set as similar users by an extraction method based on some other similarity between users.

  Next, the similar user related information extracting unit 267 acquires related information including the user ID of the similar user from the related information DB 6 based on the user ID of the similar user extracted by the similar user extracting unit 266 in step S140. Then, the similar user related information extraction unit 267 causes the user related information extraction condition determination unit 261 to select the third extraction condition, and selects the third extraction condition selected from the related information including the acquired similar user ID. The related information that satisfies is extracted as similar user related information (that is, items previously purchased by similar users are extracted) (step S150).

  Next, the recommendation index calculation unit 268 extracts the item IDs included in the similar user related information extracted by the similar user related information extraction unit 267 in step S150 so as not to overlap. Then, the recommendation index calculation unit 268 calculates a recommendation index based on the extracted item ID (step S160). Here, with reference to FIG. 14 again, the process of calculating the recommendation index will be described. As shown in the column C2 of FIG. 14, the recommendation index calculation unit 268 selects the extracted item IDs one by one, and calculates the similarity between users of all similar users who have purchased the selected item IDs in the past. The recommendation index idx is calculated by summing up.

  In column C2, item IDi4 has been purchased in the past by similar users indicated by user IDs u2 and u5. Therefore, the recommendation index idx (i4) of the item indicated by the item ID i4 is 0.70, which is the sum of the similarities between the similar users sim (u2) and sim (u5) indicated by the user IDs u2 and u5. As with the item IDi4, the recommendation index idx (i5) of the item indicated by the item IDi5 is 0.90, and the recommendation index idx (i6) of the item indicated by the item IDi6 is 0.70.

  Next, the recommended item extraction unit 269 extracts an item ID indicating a recommended item from the item ID extracted in step S150 based on the recommendation index calculated in step S160 (step S170). Here, a process in which the recommended item extracting unit 269 extracts the recommended item will be described. The recommended item extraction unit 269 extracts a predetermined number of item IDs as item IDs indicating recommended items in descending order of the recommendation index calculated in step S160. In this example, the predetermined number is 3, but instead, any number may be used as long as it is an integer of 1 or more.

  Instead of this, the recommended item extraction unit 269 may be configured to extract an item ID having a recommended index higher than a predetermined value as an item ID indicating a recommended item, and the item ID of the item ID for which the recommended index is calculated. It may be configured to extract all as item IDs indicating recommended items, or may be configured to extract item IDs extracted by extraction methods based on other recommended indexes as item IDs indicating recommended items. .

  In the example illustrated in FIG. 14, the recommended item extraction unit 269 extracts three items indicated by the item IDs i4 to i6 as recommended items. Note that, as shown in the column C3, the items indicated by the item IDi5 have the highest relationship, and the items indicated by the item IDi4 and i6 have the same value. Therefore, if the recommended item extraction unit 269 is configured to extract the item indicated by the item ID of the highest recommendation index as the recommended item, the recommended item extraction unit 269 extracts only the item indicated by the item ID i5 as the recommended item. .

  Hereinafter, a process of extracting specific related information based on the extraction condition will be described with reference to FIG. FIG. 15 is a flowchart illustrating an example of a process flow for extracting specific related information based on the extraction condition. The process shown in the flowchart of FIG. 15 includes a first extraction process by the target user related information extraction unit 262, a second extraction process by the similar candidate user related information extraction unit 264, and a third extraction by the similar candidate user related information extraction unit 264. Each process is performed in the same flow. Each difference is that in the case of the first extraction process, the extraction condition is the first extraction condition, in the case of the second extraction process, the extraction condition is the second extraction condition, and in the case of the third extraction process, the extraction condition is This is the third extraction condition.

  Therefore, in the following, for convenience of explanation, the target user related information extraction unit 262, the similar candidate user related information extraction unit 264, and the similar user related information extraction unit 267 are collectively referred to as a related information extraction unit unless it is necessary to distinguish between them. explain. Unless there is a need for distinction, the first extraction process, the second extraction process, and the third extraction process will be collectively referred to as an extraction process. Unless there is a need for distinction, the first extraction condition, the second extraction condition, and the third extraction condition will be collectively referred to as an extraction condition. In the following, for convenience of explanation, related information extracted by the flowchart shown in FIG. 15 (that is, target user related information, similar candidate user related information, similar user related information) is collectively referred to as extracted related information. explain.

  First, the related information extraction unit causes the user related information extraction condition determination unit 261 to select an extraction condition, and acquires the selected extraction condition (step S200). Below, as an example, it demonstrates on the assumption that three extraction conditions, the extraction condition which concerns on the number of extraction, the extraction condition which concerns on a date, and the extraction condition which concerns on a category were selected as an extraction condition. Next, the related information extraction unit acquires related information from the related information DB 6 (step S210).

  Here, when the related information extracting unit is the target user related information extracting unit 262, the target user related information is acquired from the related information DB 6 in step S210. If the related information extraction unit is the similar candidate user related information extraction unit 264, similar candidate user related information is acquired from the related information DB 6 in step S210. If the related information extraction unit is the similar user related information extraction unit 267, similar user related information is acquired from the related information DB 6 in step S210.

  Next, the related information extraction unit selects the related information acquired in step S210 one by one, and repeats the processing from step S230 to step S260 for the selected related information, thereby acquiring the extracted Related information that satisfies the conditions is extracted as extracted related information. At this time, the related information extraction unit repeats the processing from step S230 to step S260, and the number of extraction related information is less than the specified number specified by the extraction condition related to the number of extractions among the extracted extraction conditions. The process is repeated as long as the number i of repetitions is equal to or less than the number of related information acquired in step S210 (step S220).

  If it is determined in step S220 that the iterative process is to be performed (step S220—Yes), the related information extraction unit specifies that the selected related information is specified by the extraction condition related to the date and time among the acquired extraction conditions. It is determined whether the information is related information related to an item purchased within the date and time (step S230).

  When it is determined that the selected related information is not related information related to an item purchased within the specified date and time specified by the extraction condition related to the date and time among the acquired extraction conditions (No in step S230) The information extraction unit returns to step S220 and selects the next related information. On the other hand, when it is determined that the selected related information is related information related to an item purchased within the specified date and time specified by the extraction condition related to the date and time among the acquired extraction conditions (Yes in step S230) The related information extraction unit acquires item information of the item indicated by the item ID included in the selected related information from the item information DB 5 (step S240).

  Next, the related information extraction unit extracts the category of the item indicated by the item ID included in the selected related information based on the item information acquired in step S240, according to the category of the acquired extraction conditions. It is determined whether or not the specified category is specified by the condition (step S250).

  When it determines with it not being the designated category designated by the extraction condition which concerns on the category among the acquired extraction conditions (step S250-No), a related information extraction part returns to step S220 and selects the next related information. . On the other hand, when it determines with it being the designated category designated by the extraction condition which concerns on the category among the acquired extraction conditions (step S250-Yes), a related information extraction part uses the selected related information as extraction related information. Extract (step S260).

  Hereinafter, with reference to FIGS. 16 to 18, selection of recommended items by the control unit 26 when the combination of the first extraction condition, the second extraction condition, and the third extraction condition is different will be described. Note that the example shown in FIG. 14 is an example in which each of the first extraction condition, the second extraction condition, and the third extraction condition is the total extraction condition. FIG. 16 shows an example of similarity between users, a recommendation index, and a recommendation item when a condition for extracting related information after 2013 (an example of an extraction condition related to date and time) is selected as the first extraction condition. FIG. Note that the example shown in FIG. 16 is an example based on the user information, item information, and related information shown in FIGS. In the example illustrated in FIG. 16, the second extraction condition and the third extraction condition are described as being all extraction conditions.

  Table T-ui2 (a table representing information extracted based on the first extraction condition) is that each of the six users indicated by the user IDs u1 to u6 has purchased each of the items indicated by the item IDs i1 to i6. In this case, the information indicates whether or not an item has been purchased by the user by making a blank in the case of not purchasing and leaving it blank in the case of not purchasing. In FIG. 16, it is assumed that the user indicated by the user ID u1 is the target user.

  In the example illustrated in FIG. 16, the target user related information extraction unit 262 performs the first extraction process based on the extraction condition related to the date and time according to the first extraction condition, and thus the table T-ui2 is the table T illustrated in FIG. Compared with -ui1, information indicating whether or not an item has been purchased by the user (that is, the arrangement of circles in table T-ui2) is different. This is because the preference of the target user after 2013 is reflected by the first extraction process.

  The column C4 shows the similarity between users calculated based on information indicating whether or not an item has been purchased by the user represented by the table T-ui2. In the column C5, a recommendation index calculated based on the similarity between users shown in the column C4 is shown. In the example shown in FIG. 16, the recommended index related to the recommended item has the highest recommended index related to item IDi6, the next highest recommended index related to item IDi5, and the item IDi4 as shown in column C6. The recommendation index concerning is the lowest.

  The magnitude relationship of the recommendation index related to the recommendation item shown in FIG. 16 reflects the result of extracting the related information by the extraction condition related to the date and time by the first extraction process. That is, this indicates that the magnitude relation of the recommendation index related to the recommendation item shown in FIG. 16 reflects the preference of the target user.

  On the other hand, FIG. 17 shows the similarity between users, the recommendation index, and the recommendation item when an extraction condition for extracting related information after 2013 (an example of an extraction condition related to date and time) is selected as the second extraction condition. It is a figure which shows an example. Note that the example shown in FIG. 17 is an example based on the user information, item information, and related information shown in FIGS. In the example illustrated in FIG. 17, the first extraction condition and the third extraction condition are assumed to be all extraction conditions.

  Table T-ui3 (a table representing information extracted based on the second extraction condition) is that each of the six users indicated by the user IDs u1 to u6 has purchased each of the items indicated by the item IDs i1 to i6. In this case, the information indicates whether or not an item has been purchased by the user by making a blank in the case of not purchasing and leaving it blank in the case of not purchasing. In FIG. 17, it is assumed that the user indicated by the user ID ui1 is the target user.

  In the example illustrated in FIG. 17, the similar candidate user related information extraction unit 264 performs the second extraction process based on the extraction condition related to the date and time according to the second extraction condition, and thus the table T-ui3 is the table illustrated in FIG. 14. Compared with T-ui1 and table T-ui2 shown in FIG. 16, information indicating whether or not an item has been purchased by the user (that is, arrangement of circles in table T-ui3) is different. This is because the preference of the similar candidate user after 2013 is reflected by the second extraction process.

  The column C7 shows the similarity between users calculated based on information indicating whether or not an item has been purchased by the user represented by the table T-ui3. In the column C8, a recommendation index calculated based on the similarity between users shown in the column C7 is shown. In the example shown in FIG. 17, as shown in the column C9, the recommendation index related to the recommended item has the highest recommendation index related to the item IDi4, the next highest recommended index related to the item IDi5, and the item IDi6. The recommendation index concerning is the lowest.

  The magnitude relation of the recommendation index related to the recommendation item shown in FIG. 17 reflects the result of extracting the related information by the extraction condition related to the date and time by the second extraction process. In other words, this indicates that the preference index related to the recommendation item shown in FIG. 17 reflects the preference of the similar candidate user. The magnitude relationship of the recommendation index related to the recommended item reflecting the preference of the similar candidate user is, as can be seen by comparing FIG. 16 and FIG. 17, the magnitude relationship of the recommendation index related to the recommended item reflecting the preference of the target user. It becomes a different relationship.

  On the other hand, FIG. 18 shows that the extraction condition for extracting the category c1 (an example of the extraction condition related to the category) is selected as the first extraction condition and the second extraction condition, and the extraction condition for extracting the category c2 (in the category The user when the related information of a specific category is extracted by each of the first extraction process, the second extraction process, and the third extraction process by selecting another example of the extraction condition as the third extraction condition. It is a figure which shows an example of between similarity, a recommendation index | exponent, and a recommendation item. The example shown in FIG. 18 is an example based on the user information, item information, and related information shown in FIGS.

  The table T-ui4 (a table representing information extracted based on the first extraction condition and the second extraction condition) and the table T-ui5 (a table representing information extracted based on the third extraction condition) include a user ID u1. Each of the six users indicated by -u6 indicates whether or not an item has been purchased by the user by purchasing each item indicated by the item IDs i1 to i6, and leaving it blank if not purchased. Represents information. In FIG. 18, it is assumed that the user indicated by the user ID ui1 is the target user.

  In the example illustrated in FIG. 18, the target user related information extraction unit 262 performs the first extraction process based on the extraction condition related to the category according to the first extraction condition, and the similar candidate user related information extraction unit 264 performs the first extraction process based on the extraction condition related to the category. Since the second extraction process is performed and the third extraction process by the similar user related information extraction unit 267 is performed, the table T-ui4 and the table T-ui5 are the same as the table T-ui1 illustrated in FIG. Compared with the table T-ui2 and the table T-ui3 shown in FIG. 17, the information indicating whether or not the user has purchased an item (that is, the arrangement of the circles in the table T-ui4) is different. This is because the preference of the category of the target user is reflected by the first extraction process, the preference of the category of the similar candidate user is reflected by the second extraction process, and the preference of the category of the similar user is reflected by the third extraction process. Is the cause.

  The column C10 shows the similarity between users calculated based on information indicating whether or not an item has been purchased by the user represented by the table T-ui4 and the table T-ui5. In the column C11, a recommendation index calculated based on the similarity between users shown in the column C10 is shown. In the example shown in FIG. 18, the recommended index related to the recommended item has the highest recommended index related to the item IDi4 and the lowest recommended index related to the item IDi6 as shown in the column C12.

  The magnitude relationship of the recommendation index related to the recommendation item shown in FIG. 18 reflects the result of extracting the related information according to the extraction condition related to the category by each of the first extraction process, the second extraction process, and the third extraction process. . That is, this indicates that the magnitude relation of the recommendation index related to the recommendation item shown in FIG. 18 reflects all preferences of the target user, the similar candidate user, and the similar user. The magnitude relation of the recommendation index related to the recommended item reflecting all the preferences of the target user, the similar candidate user, and the similar user is larger or smaller than that shown in FIGS. The relationship is different.

  In addition, as illustrated in FIG. 18, by performing extraction of related information based on the extraction condition related to the category by either one or both of the second extraction process and the third extraction process, the information providing apparatus 2 Items in categories that the user has not purchased in the past can also be presented, and a higher sales promotion effect can be produced. In addition, the information providing device 2 extracts, for example, related information of the category “travel” by the first extraction process, and extracts related information of the category “meal” by the second extraction process. Recommendation items related to various categories can be presented to the target user, and as a result, more flexible information provision can be performed.

  As illustrated in FIGS. 16 to 18, in the information providing apparatus 2, the user-related information extraction condition determination unit 261 of the control unit 26 is different for each of the first extraction process, the second extraction process, and the third extraction process. Since the information processing terminal 3 requests the information providing device 2 to display the recommended item, the user can be presented with a different recommended item to select the extraction condition. In other words, the information providing apparatus 2 is an item that matches the target user's preference to the target user, and can present more various items. As a result, the information providing device 2 can enhance the sales promotion effect.

  As described above, the information providing system 1 according to the first embodiment extracts the related information related to the item that satisfies the first extraction condition among the items purchased by the target user as the target user related information, and the extracted target The user who purchased the item included in the user related information is extracted as a similar candidate user, the related information related to the item satisfying the second extraction condition among the items purchased by the similar candidate user is extracted as the similar candidate user related information, and the target Based on user-related information and similar candidate user-related information, a user similar to the target user is extracted as a similar user from among similar candidate users, and related items that satisfy the third extraction condition among items purchased by similar users Information is extracted as similar user related information. Thereby, the information provision system 1 can perform various information extraction.

Second Embodiment
Hereinafter, a second embodiment of the present invention will be described with reference to the drawings. The information providing system 1 according to the second embodiment includes an information providing apparatus 2 a instead of the configuration including the information providing apparatus 2. FIG. 19 is a diagram illustrating an example of a functional configuration of the information providing apparatus 2a according to the second embodiment. Note that in the second embodiment, the same components as those in the first embodiment are denoted by the same reference numerals, and description thereof is omitted. The information providing apparatus 2a according to the second embodiment has an item ID indicating an item input to the information processing terminal 3 by the target user, instead of a configuration in which the recommended item is selected based on the user ID included in the login information of the target user. And based on the item ID output to the information provision apparatus 2a from the information processing terminal 3, a recommended item is selected.

The information providing apparatus 2a includes a storage unit 22, an input receiving unit 23, a communication unit 25, and a control unit 26a.
The control unit 26a includes an item related information extraction condition determination unit 271, a target item related information extraction unit 272, a similar candidate item extraction unit 273, a similar candidate item related information extraction unit 274, and an item similarity calculation unit 275. , A similar item extraction unit 276 and a recommended item output unit 270.

  In response to a request from the information processing terminal 3, the control unit 26 a outputs various information related to the website provided by the information providing apparatus 2 to the information processing terminal 3 via the communication unit 25. Further, the control unit 26 a acquires an item ID from the information processing terminal 3. Hereinafter, the item indicated by the item ID acquired by the control unit 26a from the information processing terminal 3 will be referred to as a target item and described. The control unit 26a is an example of an acquisition unit.

  The item related information extraction condition determination unit 271 sets a plurality of item related information extraction conditions stored in the storage unit 22 in response to requests from the target item related information extraction unit 272 and the similar candidate item related information extraction unit 274, respectively. Read. The item related information extraction condition is related information stored in the related information DB 6, and the desired related information is obtained from the related information read by each of the target item related information extracting unit 272 and the similar candidate item related information extracting unit 274. This is a condition for extraction.

  For example, a plurality of item related information extraction conditions include an extraction condition related to date and time, an extraction condition related to gender (for example, a condition for extracting related information related to an item purchased by a man or a woman), and an extraction related to age Conditions (for example, conditions for extracting related information related to items purchased by people in their 20s), extraction conditions related to the number of extractions (for example, to extract related information related to the 10 items purchased most recently) And all extraction conditions are included. In the following, for convenience of explanation, unless it is necessary to distinguish other item-related information extraction conditions (for example, the above-described extraction conditions related to gender, extraction conditions related to age, etc.) other than all extraction conditions, they are referred to as specific extraction conditions. I will explain.

  The item related information extraction condition determination unit 271 selects the read item related information extraction condition according to a predetermined selection rule. The predetermined selection rule is, for example, selecting at random from a plurality of item related information extraction conditions, selecting in a predetermined order, or the like. Further, when selecting the item related information extraction condition, the item related information extraction condition determination unit 271 selects a different number of item related information extraction conditions from the plurality of item related information extraction conditions each time the item related information extraction condition is selected.

  Note that the item related information extraction condition determination unit 271 always extracts the same number of item related information extractions instead of a configuration in which a different number of item related information extraction conditions are selected from the plurality of item related information extraction conditions each time selection is made. A configuration for selecting conditions may be used, or a configuration for periodically changing the number to be selected may be used. The item-related information extraction condition determination unit 271 outputs the selected item-related information extraction condition to the requesting function unit. The item related information extraction condition determination unit 271 is an example of an extraction condition determination unit.

  The target item related information extraction unit 272 acquires related information associated with the item ID from the related information DB 6 based on the item ID of the target item. Then, the target item related information extraction unit 272 causes the item related information extraction condition determination unit 271 to select item related information extraction conditions. Hereinafter, for convenience of explanation, the item related information extraction condition selected by the item related information extraction condition determination unit 271 in response to a request from the target item related information extraction unit 272 will be referred to as a fourth extraction condition.

  The target item related information extraction unit 272 acquires the fourth extraction condition selected by the item related information extraction condition determination unit 271 and is acquired from the related information including the item ID of the target item acquired from the related information DB 6. Related information that satisfies the fourth extraction condition is extracted as target item related information. Hereinafter, for convenience of explanation, the process of extracting the target item related information will be referred to as a fourth extraction process. For example, when the fourth extraction condition is a condition for extracting related information related to a male, the target item related information extracting unit 272 selects related information related to a male user from among users who purchased the target item. Extract as related information.

  In addition, for example, if the fourth extraction condition is a condition for extracting related information related to a man and a condition for extracting related information related to a 20s, the target item related information extracting unit 272 The related information related to the male user in his 20s is extracted as the target item related information.

  Thus, the target item related information extraction unit 272 can extract the target item related information according to the fourth extraction condition selected by the item related information extraction condition determination unit 271. Accordingly, the target item related information extraction unit 272 can extract different target item related information each time its own function is executed. The target item related information extraction unit 272 outputs the extracted target item related information to the similar candidate item extraction unit 273 and the item similarity calculation unit 275.

  When the similar item extraction unit 273 acquires the target item related information from the target item related information extraction unit 272, the similar candidate item extraction unit 273 refers to the acquired target item related information, and there is no overlap with the users included in the target item related information as target users. Extract as follows. And the similar candidate item extraction part 273 acquires the related information which concerns on the item which the target user extracted in the past has purchased from related information DB6 as similar candidate item related information. The similar candidate item extraction unit 273 extracts items included in the acquired similar candidate item related information as similar candidate items so as not to overlap.

  By this extraction, the similar candidate item extraction unit 273 can extract an item purchased by a user who is highly likely to have a similar preference as the user who purchased the target item as a similar candidate item. The similar candidate item extraction unit 273 outputs the item ID indicating the extracted similar candidate item to the similar candidate item related information extraction unit 274 together with the similar candidate item related information.

  The similar candidate item related information extraction unit 274 acquires an item ID indicating the similar candidate item together with the similar candidate item related information. The similar candidate item related information extraction unit 274 acquires, from the related information DB 6, related information related to a user who has purchased the acquired item ID in the past. Then, the similar candidate item related information extraction unit 274 causes the item related information extraction condition determination unit 271 to select an item related information extraction condition. Hereinafter, for convenience of explanation, the item related information extraction condition selected by the item related information extraction condition determination unit 271 in response to a request from the similar candidate item related information extraction unit 274 will be referred to as a fifth extraction condition.

  The similar candidate item related information extraction unit 274 acquires the fifth extraction condition selected by the item related information extraction condition determination unit 271 and is acquired from the related information related to the user who purchased the acquired item ID in the past. Related information that satisfies the fifth extraction condition is extracted as similar candidate item related information. Hereinafter, for convenience of explanation, the process of extracting the fifth related information will be referred to as a fifth extraction process. For example, when the fifth extraction condition is a condition for extracting related information related to a man, the similar candidate item related information extracting unit 274 displays related information related to a male user who purchased the similar candidate item. Extract as information.

  In addition, the similar candidate item related information extraction unit 274 is, for example, a condition for causing the fifth extraction condition to extract related information related to a male, and for extracting related information related to a user who has purchased an item in his 20s. If the condition is satisfied, the related information related to the male user in his 20s who purchased the similar candidate item is extracted as the similar candidate item related information. As described above, the similar candidate item related information extraction unit 274 can extract similar candidate item related information according to the fifth extraction condition selected by the item related information extraction condition determination unit 271. As a result, the similar candidate item related information extracting unit 274 can extract different similar candidate item related information each time its own function is executed. The similar candidate item related information extraction unit 274 outputs the extracted similar candidate item related information to the inter-item similarity calculation unit 275.

  The inter-item similarity calculation unit 275 acquires the target item related information from the target item related information extraction unit 272. Further, the inter-item similarity calculation unit 275 acquires similar candidate item related information from the similar candidate item related information extraction unit 274. Based on the acquired target item related information and similar candidate item related information, the inter-item similarity calculation unit 275 calculates an inter-item similarity that represents the similarity with the target item for each similar candidate item. The item similarity calculation unit 275 outputs the calculated item similarity to the similar item extraction unit 276.

  The similar item extraction unit 276 acquires the item similarity from the item similarity calculation unit 275. The similar item extraction unit 276 converts (replaces) the acquired similarity between items into a recommendation index, and extracts an item whose replaced recommendation index is a predetermined value or more as a recommended item (similar item). The similar item extraction unit 276 may be configured to extract a recommended item by some other method based on the recommendation index, instead of extracting an item having a recommendation index of a predetermined value or more as a recommended item. The similar item extraction unit 276 outputs an item ID indicating the extracted recommendation item to the recommendation item output unit 270.

  Hereinafter, with reference to FIG. 20, the recommendation screen which the information provision apparatus 2a displays on the information processing terminal 3 is demonstrated. FIG. 20 is a diagram illustrating an example of the recommendation screen P <b> 2 displayed on the information processing terminal 3. On the recommendation screen P2, the target item selected by the target user is displayed in the region R4, and the recommendation item selected by the information providing apparatus 2a based on the target item is displayed in the region R5.

  Note that the recommendation screen P2 shown in FIG. 20 is merely an example. Therefore, the recommendation item displayed on the recommendation screen P2 may be classified into some category. In FIG. 20, the recommended items included in the region R5 are not arranged in the ranking order, but may be displayed in any ranking order. In FIG. 20, thumbnails are displayed as information indicating recommended items, but instead, moving images related to recommended items may be displayed.

  The target user selects an item that the target user is interested in on the website without performing a spontaneous search by displaying the recommendation screen P2 as shown in FIG. 20 on the information processing terminal 3. The item (that is, the recommended item) that is likely to be similar to the preference related to the item selected by itself can be easily found. In other words, the information providing apparatus 2a can present to the target user recommended items that are likely to meet the target user's preference, and as a result, sales of items on a website or the like that sells the recommended items. Can be promoted.

  Hereinafter, a flow until a recommended item is extracted by the information providing apparatus 2a will be described with reference to FIG. FIG. 21 is a diagram schematically illustrating an example of a flow until a recommended item is extracted by the information providing apparatus 2a. In the following, for convenience of explanation, processing by each functional unit of the control unit 26a will be described as processing of the control unit 26a. However, actually, the processing is performed by each functional unit described with reference to FIG. In addition, the example illustrated in FIG. 21 is an example where both the fourth extraction condition and the fifth extraction condition are specific extraction conditions. Note that one or both of the fourth extraction condition and the fifth extraction condition may be all extraction conditions.

  First, the control unit 26a performs a fourth extraction process based on the fourth extraction condition based on the item ID of the target item I17, so that the user who satisfies the fourth extraction condition and purchased the target item I17 in the past. The user (currently owned) is extracted as the target user U13 (step S41). Here, it is assumed that the fourth extraction condition in the example illustrated in FIG. 21 is a condition for extracting a male user. That is, the user who satisfies the fourth extraction condition is a man who purchased the target item I17 in the past. Based on the fourth extraction condition, the control unit 26a extracts the male who purchased the target item I17 in the past as the target user U13.

  Next, the control unit 26a extracts the item purchased in the past by the target user U13 extracted in step S41 as the similar candidate item I18 (step S42). Next, the control unit 26a performs a fifth extraction process based on the fifth extraction condition for each similar candidate item I18, so that the user who satisfies the fifth extraction condition and has purchased the similar candidate item I18 in the past. The user is extracted as a similar candidate user U14 (step S43).

  Here, it is assumed that the fifth extraction condition in the example shown in FIG. 21 is a condition for extracting a user in their 20s. That is, the user who satisfies the fifth extraction condition is a user in his 20s who has purchased the similar candidate item I18 in the past. For each similar candidate item I18, the control unit 26a extracts, as the similar candidate user U14, the target user U13 in the 20s who purchased the similar candidate item I18 in the past based on the fifth extraction condition. Here, the similar candidate user U14 extracted in step S43 reflects the influence of the fifth extraction process in step S43, and becomes a user in his twenties.

  Next, the control unit 26a calculates the inter-item similarity based on the target user U13 for each target item I17 extracted in step S41 and the similar candidate user U14 for each similar candidate item I18 extracted in step S42. (Step S44). Then, the control unit 26a replaces the similarity between items calculated in step S44 with a recommendation index, and extracts (selects) the recommended item I19 (similar item I19) from the similar candidate item I18 (step S45). For example, the control unit 26a extracts (selects) a similar candidate item I18 having a recommendation index higher than a predetermined value as the recommendation item I19 (step S46).

  Instead of this, the control unit 26a may be configured to extract the recommendation item I19 by some other method based on the recommendation index. Here, the recommendation index (similarity between items) is calculated based on the relevance between the user who purchased the target item in the past and the user who purchased the similar candidate item in the past. The value reflects both the attribute of the user having the preference to try and the attribute of the user having the preference to purchase the similar candidate item.

  Therefore, the recommendation item I19 selected based on the recommendation index reflects the attribute of the user who has a preference to purchase the target item and the attribute of the user who has a preference to purchase the similar candidate item. Thus, the control unit 26a selects an extraction condition selected according to a predetermined selection rule from a plurality of specific extraction conditions and all extraction conditions for each of the fourth extraction condition and the fifth extraction condition. Thus, every time a recommended item is displayed, a different recommended item can be presented to the target user.

  Hereinafter, with reference to FIG. 22, the process of extracting the recommended item by the control unit 26a will be described. FIG. 22 is a flowchart illustrating an example of a process flow for extracting recommended items by the control unit 26a. Hereinafter, the control unit 26a will be described on the assumption that the item ID has already been acquired from the information processing terminal 3. First, the target item related information extraction unit 272 acquires related information including the item ID of the target item from the related information DB 6. Then, the target item related information extraction unit 272 causes the item related information extraction condition determination unit 271 to select the fourth extraction condition, and selects the fourth extraction condition from the related information including the item ID of the acquired target item. The target item related information is extracted based on that (that is, the user who has purchased the target item in the past is extracted) (step S300).

  Next, the similar candidate item extraction unit 273 selects items that the user (target user) indicated by the user ID included in the target item related information extracted by the target item related information extraction unit 272 has purchased in the past. The item ID of the similar candidate item is extracted as an item (step S310). Next, the similar candidate item related information extraction unit 274 acquires related information including the item ID of the similar candidate item from the related information DB 6. Then, the similar candidate item related information extraction unit 274 causes the item related information extraction condition determination unit 271 to select the fifth extraction condition, and is selected from the related information including the item ID of the acquired similar candidate item. Related information that satisfies the condition is extracted as similar candidate item related information (that is, a user who has purchased a similar candidate item in the past is extracted) (step S320).

  Next, the inter-item similarity calculation unit 275 is based on the target item related information extracted by the target item related information extraction unit 272 and the similar candidate item related information extracted by the similar candidate item related information extraction unit 274. The similarity between items between the target item and the similar candidate item is calculated (step S330). Here, with reference to FIG. 23, processing for calculating the similarity between items will be described. FIG. 23 is a diagram for explaining an example of processing for calculating the similarity between items. The example shown in FIG. 23 is an example based on the user information, item information, and related information shown in FIGS. In the example illustrated in FIG. 23, the fourth extraction condition and the fifth extraction condition will be described as being all extraction conditions.

  The table T-iu1 indicates that each of the six items indicated by the item IDs i1 to i6 is ○ when the six users indicated by the user IDs u1 to u6 have purchased in the past. By making it blank, information indicating which user has purchased for each item is represented. In FIG. 23, it is assumed that the item indicated by item IDi1 is the target item.

  When the set of users who purchased the item R is V and the set of users who purchased the item S is W, the item-to-item similarity calculation unit 275 uses the Jaccard coefficient based on the above equation (1) as the item-to-item similarity sim. calculate. The column C13 shows the similarity between items calculated based on information indicating which user has purchased for each item represented by the table T-iu1. Similar to the similarity between users, the item similarity may be a value calculated by some other method (for example, a Cosine coefficient) instead of the Jaccard coefficient.

  Next, the similar item extraction unit 276 replaces the item similarity calculated by the item similarity calculation unit 275 in step S330 as a recommendation index. And the similar item extraction part 276 extracts a recommendation item based on a recommendation index | exponent and item ID of the extracted similar item (step SS340).

  Hereinafter, with reference to FIG. 24, a process in which the control unit 26a extracts specific related information based on the extraction condition will be described. FIG. 24 is a flowchart illustrating an example of a processing flow in which the control unit 26a extracts specific related information based on the extraction condition. The process shown in the flowchart shown in FIG. 23 is performed in the same flow in each of the fourth extraction process by the target item related information extraction unit 272 and the fifth extraction process by the similar candidate item related information extraction unit 274. Each difference is that in the case of the fourth extraction process, the extraction condition is the fourth extraction condition, and in the case of the fifth extraction process, the extraction condition is the fifth extraction condition.

  Therefore, in the following, for the convenience of explanation, the target item related information extraction unit 272 and the similar candidate item related information extraction unit 274 will be collectively referred to as a related information extraction unit unless there is a need for distinction. Further, unless there is a need to distinguish, the fourth extraction process and the fifth extraction process will be collectively referred to as an extraction process. In addition, the fourth extraction condition and the fifth extraction condition will be collectively referred to as an extraction condition and described unless there is a need for distinction. Further, in the following, for convenience of explanation, related information (that is, target item related information and similar candidate item related information) extracted by the flowchart shown in FIG. 23 will be collectively referred to as extracted related information.

  First, the related information extraction unit causes the item related information extraction condition determination unit 271 to select an extraction condition, and acquires the selected extraction condition (step S400). Below, as an example, it demonstrates on the assumption that three extraction conditions, the extraction condition which concerns on the number of extraction, the extraction condition which concerns on date, the extraction condition which concerns on sex, and the extraction condition which concerns on age, were selected as an extraction condition. Next, the related information extraction unit acquires related information from the related information DB 6 (step S410).

  Here, when the related information extraction unit is the target item related information extraction unit 272, in step S410, related information including the item ID of the target item is acquired from the related information DB 6. When the related information extracting unit is the similar candidate item related information extracting unit 274, related information including the item ID of the similar candidate item is acquired from the related information DB 6 in step S410.

  Next, the related information extraction unit selects the related information acquired in step S410 one by one, and repeats the processing from step S430 to step S470 for the selected related information, thereby acquiring the extracted information. Related information that satisfies the conditions is extracted as extracted related information. At this time, the related information extraction unit repeats the processing from step S430 to step S470, and the number of extraction related information is less than the designated number specified by the extraction condition related to the number of extractions among the extracted extraction conditions. The process is repeated as long as the number i of repetitions is equal to or less than the number of related information acquired in step S410 (step S420).

  If it is determined in step S430 that the iterative process is to be performed (step S420—Yes), the related information extracting unit selects the related information within the specified date and time based on the extraction condition related to the date and time among the acquired extraction conditions. It is determined whether it is information (step S430). When it is determined that the selected related information is not related information within the designated date and time (step S430-No), the related information extraction unit returns to step S430 and selects the next related information.

  On the other hand, when it is determined that the selected related information is related information within the specified date and time (step S430-Yes), the related information extraction unit sets the user ID indicating the user included in the related information extracted in step S410. Based on this, the user information is acquired from the user information DB 4 (step S440). In the present embodiment, it is assumed that the user attribute is further associated with the user ID associated with the user name in the user information. The user attributes include, for example, the user's gender and the user's age.

  Next, the related information extraction unit is related information related to a gender user (male user in this example) specified by the gender extraction condition of the acquired extraction conditions. It is determined whether or not there is (step S450). When it determines with the selected relevant information not being the relevant information which concerns on a male user (step S450-No), a relevant information extraction part returns to step S430 and selects the next relevant information.

  On the other hand, when it is determined that the selected related information is related information related to the male user (step S450-Yes), the related information extraction unit determines that the selected related information is the age of the acquired extraction conditions. It is determined whether or not the information is related information related to a user of the age specified by the extraction condition (in this example, a user in his 20s) (step S460).

  When it is determined that the selected related information is not related information related to a user in their 20s (step S460-No), the related information extraction unit returns to step S430 and selects the next related information. On the other hand, when it is determined that the selected related information is related information related to a user in their 20s (step S460-Yes), the related information extraction unit extracts the selected related information as extracted related information ( Step S470).

  As described above, the information providing system 1 according to the second embodiment extracts the related information related to the target user satisfying the fourth extraction condition among the users who purchased the target item as the target item related information, and has been extracted. Items purchased by similar candidate users included in the target item related information are extracted as similar candidate items, and related information related to users who satisfy the fifth extraction condition among users who have purchased similar candidate items is extracted as similar candidate item related information Then, based on the target item related information and the similar candidate item related information, an item similar to the target item is extracted as a similar item from the similar candidate items. Thereby, the information provision system 1 can perform various information extraction. The fourth extraction condition is an example of a first extraction condition, and the fifth extraction condition is an example of a second extraction condition.

  The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and changes, substitutions, deletions, and the like are possible without departing from the gist of the present invention. May be.

  Further, a program for realizing the function of each unit constituting each of the information providing apparatuses 2 and 2a and the information processing terminal 3 in the above embodiment is recorded on a computer-readable recording medium and recorded on the recording medium. The information providing apparatus 2 and the information processing terminal 3 may be implemented by causing the computer system to read and execute the program thus executed. Note that the “computer system” here includes an OS (Operation System) and hardware such as peripheral devices.

Further, the “computer system” includes a homepage providing environment (or display environment) if the WWW system is used.
The “computer-readable recording medium” refers to a storage device such as a flexible disk, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Furthermore, a “computer-readable recording medium” dynamically holds a program for a short time, like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. In this case, it also includes those that hold a program for a certain period of time, such as a volatile memory inside a computer system serving as a server or client in that case. The program may be a program for realizing a part of the above-described functions, or may be a program that can realize the above-described functions in combination with a program already recorded in a computer system.

(1) One aspect of the present invention is a target user related information extraction unit that extracts, as target user related information, related information related to an item that satisfies the first extraction condition among items related to a target user, and the extracted target A similar candidate user related information extraction unit that extracts related information related to an item that satisfies the second extraction condition among items related to a user related to an item included in the user related information as similar candidate user related information, and the target user related A similar user extraction unit that extracts a user similar to the target user from among users included in the similar candidate user related information based on the information and the similar candidate user related information, and is related to the similar user Related information related to items that satisfy the third extraction condition among items And similar user related information extraction unit that extracts as a user-related information like, is an information providing apparatus comprising a.

(2) Another aspect of the present invention is the information providing apparatus according to (1), in which a user related to an item related to the target user extracted by the target user related information extraction unit is a similar candidate user A similar candidate user extraction unit that extracts as the candidate information related to the candidate, and the information extraction unit relates to an item that satisfies the second extraction condition among items related to the similar candidate user extracted by the similar candidate user extraction unit. It is an information providing device that extracts related information as similar candidate user related information.
(3) Another aspect of the present invention is the information providing apparatus according to (1) or (2), wherein each of the first extraction condition, the second extraction condition, and the third extraction condition is provided. An information providing apparatus comprising: an extraction condition determining unit that assigns one or more extraction conditions from among a plurality of extraction conditions.
(4) Another aspect of the present invention is the information providing apparatus according to (3), in which the extraction condition determining unit performs the assignment based on the input condition selection information. It is.

(5) Another aspect of the present invention is the information providing apparatus according to any one of (1) to (4), wherein the degree of similarity between the target user and the similar candidate user is between The information providing apparatus includes an inter-user similarity calculating unit that calculates a similarity, and the similar user extracting unit extracts the similar user from the similar candidate users based on the calculated inter-user similarity.
(6) Another aspect of the present invention is the information providing apparatus according to (5), in which the similar user related information extracted by the similar user related information extracting unit and the inter-user similarity calculating unit Based on the calculated similarity between users, the importance of an item related to the similar user extracted by the similar user extraction unit, and the importance when providing information to the target user A recommended index calculating unit that calculates a recommended index that is a value indicated, and a recommended item to be presented to the target user based on the recommended index calculated by the recommended index calculating unit is extracted from items related to the similar user And a recommended item extraction unit.
(7) Another aspect of the present invention is the information providing apparatus according to any one of (1) to (6), wherein the first extraction condition, the second extraction condition, and the third extraction Each of the conditions is an information providing apparatus including one or more of an extraction condition related to date and time, an extraction condition related to a category, and an extraction condition related to the number of extractions.

(8) Another aspect of the present invention is an information providing system including an information providing device and an information processing terminal, and the information providing device uses information indicating a user of the information processing terminal as target user information. An acquisition unit to acquire, a target item related information extraction unit that extracts, as target user related information, related information related to an item that satisfies the first extraction condition among items related to an item related to the target user indicated by the target user information; A similar candidate user related information extraction unit that extracts related information related to an item that satisfies the second extraction condition among items related to a user related to an item included in the extracted target user related information as similar candidate user related information And the similar candidate based on the target user related information and the similar candidate user related information A similar user extraction unit that extracts a user similar to the target user as a similar user from among users, and extracts related information related to an item that satisfies the third extraction condition among items related to the similar user as similar user related information A similar user related information extracting unit, a recommended item extracting unit for extracting an item based on the extracted similar user related information as a recommended item, and a first output unit for outputting information indicating the recommended item to the information processing terminal The information processing terminal includes: a second output unit that outputs the target user information to the information providing device; and a display unit that displays information indicating the recommended item acquired from the information providing device. An information providing system is provided.

(9) According to another aspect of the present invention, the second output unit that outputs the target user information to the information providing device, and the related information related to the item satisfying the first extraction condition among the items related to the target user are related to the target user. Information related to an item that satisfies the second extraction condition among items related to a user related to an item included in the extracted target user related information extracted as information is extracted as similar candidate user related information, and the target Based on the user related information and the similar candidate user related information, a user similar to the target user is extracted as a similar user from the similar candidate users, and a third extraction condition among the items related to the similar user is satisfied. Related information related to the item is extracted as similar user related information, the extracted Similar the items based on the user information is extracted as recommended item, a display unit for displaying information indicating the recommended item extracted, an information processing terminal equipped with.

(10) In another aspect of the present invention, related information relating to an item that satisfies the first extraction condition among items related to the target user is extracted as target user related information, and is included in the extracted target user related information The related information related to the item satisfying the second extraction condition among the items related to the user related to the item is extracted as similar candidate user related information, and the similar information is extracted based on the target user related information and the similar candidate user related information. Information provision that extracts a user similar to the target user from among candidate users as a similar user, and extracts related information related to an item that satisfies the third extraction condition among items related to the similar user as similar user related information Is the method.

(11) According to another aspect of the present invention, the computer of the information providing device causes the related information related to an item that satisfies the first extraction condition among the items related to the target user to be extracted as target user related information. Related information related to an item that satisfies the second extraction condition among items related to an item included in the target user related information is extracted as similar candidate user related information, and the target user related information and the similar candidate user related are extracted. Based on the information, a user similar to the target user is extracted as a similar user from the similar candidate users, and related information related to an item satisfying the third extraction condition among items related to the similar user is related to the similar user This is a program that extracts information.

DESCRIPTION OF SYMBOLS 1 Information provision system, 2, 2a Information provision apparatus, 3 Information processing terminal, 4 User information DB, 5 Item information DB, 6 Related information DB, 21, 31 CPU, 22, 32 Storage part, 23, 33 Input reception part, 24, 34 Display unit, 25, 35 Communication unit, 26, 26a, 36 Control unit, 37 Display control unit, 261 User related information extraction condition determination unit, 262 Target user related information extraction unit, 263 Similar candidate user extraction unit, 264 Similar candidate user related information extracting unit, 265 Similarity calculating unit between users, 266 Similar user extracting unit, 267 Similar user related information extracting unit, 268, 268a Recommended index calculating unit, 269 Recommended item extracting unit, 270 Recommended item output unit, 271 Item related information extraction condition determination unit, 272 Temu related information extraction section 273 similar candidate item extracting unit 274 similar candidate item-related information extraction unit 275 inter-item similarity degree calculation section, 276 similar items extractor

Claims (11)

A target user related information extracting unit that extracts related information related to an item that satisfies the first extraction condition among items related to the target user as target user related information;
A similar candidate user related information extraction unit that extracts related information related to an item satisfying the second extraction condition among items related to a user related to an item included in the extracted target user related information, as similar candidate user related information; ,
Based on the target user related information and the similar candidate user related information, a similar user extraction unit that extracts a user similar to the target user from among users included in the similar candidate user related information,
A similar user related information extraction unit that extracts related information related to an item that satisfies the third extraction condition among items related to the similar user as similar user related information;
An information providing apparatus comprising: The information providing device according to claim 1,
A similar candidate user extraction unit that extracts a user related to an item related to the target user extracted by the target user related information extraction unit as a similar candidate user;
The similar candidate user related information extraction unit extracts, as similar candidate user related information, related information related to an item satisfying a second extraction condition among items related to the similar candidate user extracted by the similar candidate user extraction unit. ,
Information providing device. The information providing device according to claim 1 or 2,
An extraction condition determination unit that assigns one or more extraction conditions from among a plurality of extraction conditions for each of the first extraction condition, the second extraction condition, and the third extraction condition;
An information providing apparatus comprising: The information providing device according to claim 3,
The extraction condition determination unit performs the assignment based on the input condition selection information.
Information providing device. The information providing device according to any one of claims 1 to 4,
An inter-user similarity calculating unit that calculates an inter-user similarity that is a similarity between the target user and a user included in the similar candidate user related information;
The similar user extraction unit extracts the similar user from users included in the similar candidate user related information based on the calculated similarity between users.
Information providing device. The information providing device according to claim 5,
The similarity extracted by the similar user extraction unit based on the similar user related information extracted by the similar user related information extraction unit and the inter-user similarity calculated by the inter-user similarity calculation unit A recommendation index calculation unit that calculates a recommendation index that is the importance of an item related to a user and is a value indicating the importance when providing information to the target user;
A recommended item extracting unit that extracts a recommended item to be provided to the target user from items related to the similar user based on the recommended index calculated by the recommended index calculating unit;
An information providing apparatus comprising: The information providing device according to any one of claims 1 to 6,
Each of the first extraction condition, the second extraction condition, and the third extraction condition includes one or more of an extraction condition related to date and time, an extraction condition related to a category, and an extraction condition related to the number of extractions.
Information providing device. An information providing system comprising an information providing device and an information processing terminal,
The information providing apparatus includes:
An acquisition unit for acquiring information indicating a user of the information processing terminal as target user information;
A target item related information extraction unit that extracts, as target user related information, related information related to an item that satisfies the first extraction condition among items related to the target user indicated by the target user information;
A similar candidate user related information extraction unit that extracts related information related to an item satisfying the second extraction condition among items related to a user related to an item included in the extracted target user related information, as similar candidate user related information; ,
Based on the target user related information and the similar candidate user related information, a similar user extraction unit that extracts a user similar to the target user from among users included in the similar candidate user related information,
A similar user related information extraction unit that extracts related information related to an item that satisfies the third extraction condition among items related to the similar user as similar user related information;
A recommended item extracting unit that extracts an item based on the extracted similar user related information as a recommended item;
A first output unit that outputs information indicating the recommended item to the information processing terminal;
With
The information processing terminal
A second output unit for outputting the target user information to the information providing device;
A display unit for displaying information indicating the recommended item acquired from the information providing device;
Comprising
Information provision system. A second output unit for outputting the target user information to the information providing device;
Among the items related to the target user, the related information related to the item satisfying the first extraction condition is extracted as the target user related information, and among the items related to the user related to the item included in the extracted target user related information Related information related to an item satisfying the second extraction condition is extracted as similar candidate user related information, and based on the target user related information and the similar candidate user related information, from among users included in the similar candidate user related information A user similar to the target user is extracted as a similar user, and related information related to an item that satisfies the third extraction condition among items related to the similar user is extracted as similar user related information and extracted. Information item is recommended item A display unit is extracted, displays information indicating a recommended item extracted as,
An information processing terminal comprising: Extract related information related to items satisfying the first extraction condition among items related to the target user as target user related information,
Extracting related information related to an item that satisfies the second extraction condition among items related to a user related to an item included in the extracted target user related information as similar candidate user related information,
Based on the target user related information and the similar candidate user related information, a user similar to the target user is extracted as a similar user from among users included in the similar candidate user related information,
Extracting related information related to an item that satisfies the third extraction condition among items related to the similar user as similar user related information;
Information provision method. In the computer of the information provision device,
The related information related to the item satisfying the first extraction condition among the items related to the target user is extracted as the target user related information,
The related information related to the item satisfying the second extraction condition among the items related to the user related to the item included in the extracted target user related information is extracted as similar candidate user related information,
Based on the target user related information and the similar candidate user related information, a user similar to the target user is extracted as a similar user from among users included in the similar candidate user related information,
Extracting related information related to an item that satisfies the third extraction condition among items related to the similar user as similar user related information;
program.

Images