Server, method for recommendation, and program

Abstract

To provide a server capable of suitably recommending items to a user.SOLUTION: A server 10 that recommends items to a user extracts first-type items satisfying a first condition from among a plurality of items in response to a request from a terminal 20, and extracts second-type items satisfying a second condition different from the first condition from among the plurality of items. The server sequentially arranges information on the extracted first-type items and information on items extracted as the second-type items that are not the first-type items, causes the arranged information to be presented on the terminal 20, and then waits for a new request from the terminal 20.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a server, a recommendation method, and a program. [Background technology]

[0002] As a technology for recommending live streaming content to a user, a method has been proposed that includes a step of generating a first subset of live streaming programs based on a first criterion and a step of generating a second subset of live streaming programs based on a second criterion different from the first criterion (for example, Patent Document 1). In the technology described in Patent Document 1, the first subset and the second subset are recommended to a user. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2023-4874 Summary of the Invention [Problem to be solved by the invention]

[0004] When recommending items to a user, it is desirable to recommend a large number of items under various conditions.

[0005] The present invention has been made in view of the above-mentioned circumstances, and aims to provide a recommendation system etc. that can suitably recommend items to users. [Means for solving the problem]

[0006] In order to achieve the above object, the server of the present invention comprises: A server that recommends items to a user, extracting, in response to a request from the terminal, a first type item that satisfies a first condition from the plurality of items, and extracting, from the plurality of items, a second type item that satisfies a second condition different from the first condition; The information on the extracted first-type items and the information on the extracted second-type items that are not the first-type items are arranged in order and presented to the terminal, and a new request from the terminal is awaited. [Effects of the Invention]

[0007] According to the present invention, items can be suitably recommended to a user. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 2 is a functional block diagram of a recommendation system and a server according to the present embodiment. [Figure 2] FIG. 10 is a diagram illustrating an example of user information. [Figure 3] FIG. 1 is a diagram illustrating an example of a hardware configuration of an information processing device that constitutes a recommendation system according to an embodiment of the present invention. [Figure 4] 10 is a flowchart illustrating an example of a recommendation list providing process. [Figure 5] 10 is a flowchart illustrating an example of a recommendation list generation process. [Figure 6] FIG. 1 is a schematic diagram of a recommendation list. [Figure 7] 10A and 10B are diagrams showing examples of display of item information and a recommendation list. [Figure 8] FIG. 10 is a diagram showing an example of display of detailed information about an item. [Figure 9] 10 is a flowchart illustrating an example of an item extraction condition determination process. [Figure 10] FIG. 10 is a diagram illustrating an example of a tendency for the remaining ratio of a recommendation list to decrease. DETAILED DESCRIPTION OF THE INVENTION

[0009] A recommendation system 100 and a server 10 according to an embodiment of the present invention will be described below with reference to the drawings. Note that the same or corresponding parts in the drawings are denoted by the same reference numerals.

[0010] (Embodiment 1) First, a first embodiment of the present invention will be described. A functional block diagram showing the functional configuration of a recommendation system 100 and a server 10 according to this embodiment. The recommendation system 100 includes the server 10 and a terminal 20. The server 10 and the terminal 20 are connected via a communication network such as the Internet or a mobile communication network, and transmit and receive information via the communication network.

[0011] The recommendation system 100 of this embodiment is a system that recommends items to a user who uses a terminal 20. For example, the server 10 generates a recommendation list personalized for the user based on the user's item browsing history, etc., and provides the generated recommendation list to the terminal 20. The recommendation list is a list that includes a predetermined number of items to be recommended to the user.

[0012] The user can view the recommendation list provided by the server 10 by performing a scrolling operation or the like on the terminal 20. Then, for example, the user can access detailed information about an item from the recommendation list, or perform operations such as purchasing or reserving the item. In this way, by providing the recommendation list to the user, it is possible to provide the user with useful information and also to encourage the user to purchase or reserve an item. Note that the user of the terminal 20 is, for example, a purchaser of an item or a shopper who is a potential purchaser.

[0013] In order to provide useful information to users through a recommendation list and to encourage them to purchase or reserve items appropriately, it is preferable that the items included in the recommendation list are useful to users and contain a large number of items. Therefore, in this embodiment, the server 10 first extracts first-class items that satisfy a first condition that is strongly personalized for the user, and arranges the extracted first-class items in order of recommendation. Next, the server 10 extracts second-class items that satisfy a second condition that is less personalized for the user than the first condition. Then, the server 10 excludes items that overlap with the first-class items from the second-class items and arranges them after the first-class items in order of recommendation.

[0014] In this way, the server 10 generates a recommendation list in which first-type items and second-type items are arranged in order. When items are extracted based on a first condition that is highly personalized for a user, the possibility of extracting items useful to the user increases. On the other hand, when items are extracted based on the first condition, the conditions are stricter than when items are extracted based on the second condition, resulting in higher calculation costs such as search time, processing load, and calculation time. Therefore, in this embodiment, by extracting items by combining a second condition that has a lower calculation cost than the first condition, it is possible to suppress calculation costs while providing a recommendation list that includes items useful to the user and contains a larger number of items. Furthermore, by doing so, it is possible to suppress an increase in the time it takes to provide a recommendation list to the user.

[0015] Returning to Fig. 1, the server 10 is configured by, for example, a communication-capable information processing device such as a personal computer (PC) or a workstation. The server 10 may be a dedicated device. Although Fig. 1 shows one server 10, for example, multiple servers 10 may be provided for each item to be stored and managed.

[0016] The server 10 functionally comprises, for example, a transmitting / receiving unit 101, a recommendation list generating unit 102, an item information DB 103, and a user information DB (Data Base) 104, as shown in FIG.

[0017] The transmitting / receiving unit 101 transmits various types of information to the terminal 20 via the communication network, and receives various types of information from the terminal 20. For example, the transmitting / receiving unit 101 receives an item acquisition request from the terminal 20, and transmits information about the item to the terminal 20 based on the request. The transmitting / receiving unit 101 also transmits a recommendation list generated by the recommendation list generation unit 102 to the terminal 20.

[0018] The recommendation list generation unit 102 extracts multiple items that are estimated to be of high interest to the user based on the attribute information for each user (age, gender, preference settings set by the user, etc.) stored in the user information DB 104 and the item browsing history, and generates a recommendation list by arranging the extracted items in order of recommendation, such as in order of the degree of match to the conditions at the time of extraction.

[0019] The recommendation list generation unit 102 includes a first-class item extraction unit 1021 and a second-class item extraction unit 1022. The first-class item extraction unit 1021 extracts, from the item information DB 103, first-class items that satisfy a first condition that is strongly personalized for the user. The first-class item extraction unit 1021 sets search conditions for items personalized for the user based on, for example, attribute information of the user of the terminal 20, item browsing history, item purchase history, history of recommended items, history of browsing detailed information from the recommendation list, and the user's interests, fields of preference, and items analyzed based on these. Then, the first-class item extraction unit 1021 extracts, as first-class items, items that match the search conditions in the top N (N is any natural number) degree. That is, the first condition is a condition that items match the search conditions that are strongly personalized for the user in the top N degree. Note that this first condition is just an example, and the first condition may be determined in advance. For example, the first condition may be a condition that is more strongly personalized to the user than the second condition described below, and may be a condition that can extract items that the user is presumed to be interested in based on a predetermined basis. Note that if there are fewer than N items that match the search condition of the first condition, the search condition of the first condition may be relaxed so that N or more items match.

[0020] The second-type item extraction unit 1022 extracts, from the item information DB 103, second-type items that satisfy a second condition, which has a lower degree of personalization for the user than the first condition. The second-type item extraction unit 1022 sets search conditions for items based on, for example, the user's attribute information or preset preference settings (such as areas of interest). The second-type item extraction unit 1022 then extracts, as second-type items, items that match the search conditions in the top n (n is any natural number) degrees of matching. That is, the second condition is a condition that matches the search conditions in the top n degrees of matching a search condition that has a lower degree of personalization for the user. Note that this second condition is just an example, and the second condition may be predetermined. For example, the second condition may be a condition that has a lower degree of personalization for the user than the first condition, and may be a condition that can extract items that are estimated to be of some interest to the user based on predetermined grounds. For example, the second-type item extraction unit 1022 may extract popular items or newly arrived items among users in general, regardless of information about a specific user. If there are fewer than n items that match the second search condition, the second search condition may be relaxed so that n or more items match. In this embodiment, information on a huge number of items is stored in item information DB 103, and basically, it is assumed that the first and second search conditions extract a number of items that is sufficiently greater than N or n.

[0021] It can also be said that the first condition is a search condition narrower than the second condition, and the second condition is a search condition broader than the first condition.

[0022] The recommendation list generation unit 102 arranges the N number of first-type items extracted by the first-type item extraction unit 1021 in order of recommendation. Then, the recommendation list generation unit 102 removes duplicate items that overlap with the first-type items from the second-type items extracted by the first-type item extraction unit 1021, and arranges the removed items after the first-type items in order of recommendation. Then, the recommendation list generation unit 102 adds second-type items (items whose degree of match with the second condition is n+1 or higher) equal to the number of the removed extracted items. Furthermore, from the added second-type items, items that overlap with the first-type items are removed and added to the recommendation list. By repeating this process until the number of second-type items reaches n, the recommendation list generation unit 102 generates a recommendation list. The number of items in the recommendation list generated in this manner is the sum of N and n. Note that if there are enough items after excluding duplicate items from the second-type items, the number of items in the recommendation list may be the sum of N and n minus the number of duplicate items without replenishing the second-type items. Also, first-type items may be excluded from all items that match the search criteria of the second condition, and then n second-type items may be extracted.

[0023] In this way, the recommendation list generation unit 102 can generate a recommendation list that is personalized for the user and includes a variety of items. The generated recommendation list is then provided to the terminal 20 by the transmission / reception unit 101. This allows the server 10 to provide the terminal 20 with a recommendation list that includes items useful to the user and includes a variety of items.

[0024] The item information DB 103 stores multiple pieces of information about specific items. In response to an item browsing request from the terminal 20 specifying a specific item or an item browsing request specifying search criteria, the transmitter / receiver 101 provides a web page listing the item information for the specified item or a web page including detailed information about the specific item. In this embodiment, an individual item web page including detailed information about a specific item is also referred to as an item page. In this way, the server 10 also functions as a web server that provides a website whose content is item information.

[0025] Then, the server 10 generates a recommendation list by the recommendation list generation unit 102 in response to a request to view an item, etc. The transmitting / receiving unit 101 provides the recommendation list generated by the recommendation list generation unit 102 to the terminal 20 together with the item information. This makes it possible to provide the user of the terminal 20 with information on items recommended to the user in addition to item information on the specified item, thereby making it possible to provide useful information to the user. Note that items recommended to the user by the recommendation list are sometimes called recommended items. Note that recommended items may be items related to the item requested to be viewed by the user, or may be items extracted from items that the server 10 can provide (for example, items unrelated to the item requested to be viewed).

[0026] As described above, in this embodiment, the server 10 provides the recommendation list in response to a request to view items from the terminal 20. However, the present invention is not limited to this, and the server 10 may provide the recommendation list to the user in response to other triggers, such as access to a specific website or a request to acquire the recommendation list itself with specified conditions, etc. Also, the server 10 may generate a recommendation list (a list of products, etc. recommended to the user) and provide it to the terminal 20 as an advertisement on a website or SNS (Social Networking Service).

[0027] An item can be any information that a user views. For example, an item can be information about products, services, used cars, delivery services, digital content such as books, manga, music, and videos, accommodations, transportation (trains, buses, airplanes, etc.), travel, tours, restaurants, recipes, real estate, beauty facilities (hair salons, estheticians, etc.), golf courses, leisure facilities such as campsites and fishing boats, sporting events, etc., or information about purchasing or reserving any of these items.

[0028] The user information DB 104 stores user information for each user as shown in Fig. 2. The user information includes a user ID for identifying the user, attribute information of the user, the user's item browsing history, the user's item purchase history, a recommendation history for the user, and a detailed browsing history from the user's recommendation list. The user information may also include other information such as login information such as a password.

[0029] The attribute information is information about the user, such as age, sex, preference settings indicating the user's fields of interest, etc. The attribute information is input by the user, for example, when registering as a member for a service (item information providing service, etc.) provided by the server 10.

[0030] The item browsing history is information indicating the viewed items for which the user viewed the item page containing the item details. For example, simply displaying an overview of an item included in a search result list or recommendation list (such as a link to an item page containing the item name and thumbnail image) does not determine that the item has been viewed. When a specific item is selected from a search result list or recommendation list and an individual item page corresponding to that item is viewed, that item is stored as a viewed item in the browsing history of the user information. Since the fact that a user has viewed an item page can be inferred to mean that the user has a high interest in that item, the items in which the user has a high interest can be inferred from this browsing history.

[0031] The item purchase history is information indicating items that a user has purchased in the past. For example, when an item is purchased from an item page, the item is stored in the purchase history along with the purchase date and time, etc. If a user purchases an item, it can be inferred that the user has a high interest in that item, and therefore, the items in which the user has a high interest can be inferred from this purchase history.

[0032] The recommendation history is information about items that were included in recommendation lists provided to users in the past. The detailed browsing history from recommendation lists is information indicating the history of access to detailed information about an item from a recommendation list, and includes information about at least the recommendation ranking of the item in the recommendation list. Access to detailed information about an item from a recommendation list indicates that the user has viewed or glanced at at least the item in the recommendation.

[0033] The user information DB 104 may also store the results of an analysis of the user's interests and preferences based on the user's item-related history information. The analysis results may include information indicating the fields that the user likes or dislikes, product categories, specific item names, etc.

[0034] In this embodiment, the server 10 is configured to include the item information DB 103 and the user information DB 104, but at least one of the item information DB 103 and the user information DB 104 may be provided outside the server 10.

[0035] The terminal 20 is, for example, an information processing device with a communication function used by a user, such as a smartphone, tablet, or PC. While one terminal 20 is shown in FIG. 1, any number of terminals 20 may be used. The terminal 20 transmits an item viewing request to the server 10 via a communication network based on a user operation. The terminal 20 then receives item information and a recommendation list corresponding to the viewing request, which are transmitted from the server 10 based on the item viewing request. The received item information and recommendation list are displayed on the screen of the terminal 20. This allows the user to view the requested item information and also view the recommendation list, which is a list of items recommended based on the request.

[0036] Furthermore, in the recommendation system 100 of this embodiment, for example, on the terminal 20, a portion of the recommendation list is displayed on the screen. When a movement operation such as a scroll operation on the recommendation list, a movement cursor operation, or a transition operation to the next page is performed, the display range is moved so that another portion of the list is displayed on the screen. That is, the recommendation list can be displayed in a scroll display, a carousel display, or split display across multiple pages, and items in the list can be displayed one by one. Note that the display method of the recommendation list on the terminal 20 is not limited to this, and any method may be used as long as the recommendation list can be displayed appropriately.

[0037] Furthermore, when a selection operation is performed to select an item included in the recommendation list, the terminal 20 transmits a request for an item page related to that item to the server 10. The terminal 20 receives the item page corresponding to the request from the server 10 and displays the received item page on the screen. That is, the terminal 20 receives a response corresponding to the request from the server 10 and displays the item page specified in the received response on the screen.

[0038] 3 is a diagram showing an example of the hardware configuration of an information processing device constituting the server 10 and the terminal 20. The server 10 has a processor 11 such as a CPU (Central Processing Unit) or a GPU (Graphical Processing Unit), a storage device 12 such as a memory, an HDD (Hard Disk Drive) and / or an SSD (Solid State Drive), a communication interface 13 for wired or wireless communication, an input device 14 for accepting input operations, and an output device 15 for outputting information. The input device 14 is, for example, a keyboard, a touch panel, a camera, a mouse, and / or a microphone. The output device 15 is, for example, a display and / or a speaker.

[0039] The processor 11 of the server 10 reads the programs stored in the storage device 12, and operates as the various functional units shown in FIG.

[0040] The storage device 12 of the server 10 functions as an item information DB 103 and stores information on multiple items to be managed. The storage device 12 also functions as a user information DB 104 and stores user information for each user accessed by the terminal 20. At least a portion of the information on the users may be stored in the storage device 22 of the terminal 20.

[0041] Similarly, the terminal 20 has a processor 21, a storage device 22, a communication interface 23, an input device 24, and an output device 25. Note that unnecessary components in the server 10 and the terminal 20 may be omitted as appropriate, and other components may be added.

[0042] The processor 21 of the terminal 20 reads a program stored in the storage device 22 to implement a function for using the recommendation system 100. The function may be implemented by a dedicated application or a general-purpose web browser. For example, the processor 21 transmits a request to view an item or a request to obtain an item page to the server 10 in response to a user's item selection operation, and obtains item information and a recommendation list from the server 10.

[0043] Next, a description will be given of the operation of the recommendation system 100 and the server 10. Fig. 4 is a flowchart showing an example of a recommendation list providing process executed by the server 10. The recommendation list providing process is executed by the processor 11 of the server 10 reading a program stored in the storage device 12.

[0044] In the recommendation list providing process, the server 10 first determines whether or not a user request to view an item related to the server 10 has been received from the terminal 20 (step S11). The user request to view an item is transmitted in response to, for example, an item viewing operation performed by the terminal 20. If the user request to view an item has not been received (step S11; No), the recommendation list providing process is terminated.

[0045] If an item viewing request has been received (step S11; Yes), the server 10 executes a recommendation list generation process for generating a recommendation list to be provided to the terminal 20 of the user who has made the item viewing request (step S12).

[0046] 5 is a flowchart showing an example of a recommendation list generation process. The recommendation list generation process is executed by the recommendation list generation unit 102 of the server 10. In the recommendation list generation process, first, the recommendation list generation unit 102 acquires user information corresponding to the user of the terminal 20 from the user information DB 104 (step S121). In step S121, based on a request to view an item, etc., the recommendation list generation unit 102 identifies the user ID of the user of the terminal 20 who sent the request to view the item, and acquires the user information from the user information DB 104.

[0047] Then, the recommendation list generation unit 102 sets a first condition based on the acquired user information, and extracts N first-type items that match the first condition from the item information DB 103 (step S122). Note that the first condition may take into account items related to the item viewing request. For example, items similar to the items related to the user's item viewing request may be extracted, or the same category as the items related to the viewing request may be set as search conditions.

[0048] Next, the recommendation list generating unit 102 sets a second condition based on the acquired user information, and extracts second-type items that match the second condition from the item information DB 103 (step S123).

[0049] Then, the recommendation list generation unit 102 removes items that overlap with the first-type items from the extracted second-type items (step S124). In this manner, in this embodiment, the server 10 removes information about the first-type items from the extracted second-type items, and then transmits information about the second-type items that are not first-type items to the terminal 20. This makes it possible to provide information about a wide variety of items while avoiding overlapping items in the recommendation list.

[0050] The recommendation list generation unit 102 then arranges N first-type items in order of recommendation, and generates a recommendation list in which n second-type items, excluding items that overlap with the first-type items, are arranged in order of recommendation from the N+1th item onwards (step S125). The recommendation order may be based on the degree of match with each condition, etc. The recommendation list generation unit 102 then ends the recommendation list generation process.

[0051] Fig. 6 is a schematic diagram of a recommendation list generated in this way. As shown in Fig. 6, the recommendation list is composed of N (e.g., 200) first-type items and n (e.g., 160) second-type items arranged in order. In the example shown in Fig. 6, the recommendation list includes 360 items, and the items are arranged in order of recommendation from 1st to 360th.

[0052] When a user browses a web page with many items, such as a recommendation list, the percentage of time the user remains on the web page, or the percentage of time the user continues to browse or glance at the web page, tends to decrease depending on the number of items glanced at by the user. The number N of items extracted as first-type items may be predetermined based on, for example, statistics of the percentage of time the user remains on the recommendation list. For example, N may be set based on the average or median of the items glanced at by the user in the recommendation list provided to the user. Similarly, the number n of items extracted as second-type items may be predetermined based on statistics of the percentage of time the user remains on the recommendation list. For example, n may be set based on the maximum number of items glanced at by the user in the recommendation list provided to the user. In the first embodiment, N is predetermined to be 200 and n is predetermined to be 160.

[0053] By providing information on type 1 items that are highly personalized to the user up to N items where a certain percentage of items are expected to remain on the user's recommendation list, it is possible to provide the user with useful information and encourage them to access detailed information on type 1 items and purchase or reserve type 1 items. Furthermore, by providing information on type 2 items that are less personalized to the user after N items where the percentage of items remaining on the user's recommendation list decreases, it is possible to reduce calculation costs and provide information on a sufficient number of items to users who continue to glance at the recommendation list.

[0054] Upon completion of the recommendation list generation process, returning to FIG. 4, the server 10 provides the recommendation list generated in step S12 to the user terminal 20 together with item information corresponding to the viewing request (step S13). The recommendation list provided in this manner is presented to the user terminal 20 together with the item information. If the viewing request is a request to view an item page specifying a specific item, the item information corresponding to the viewing request may be an item page including detailed information about the specific item. If the viewing request is a request to view an item page specifying search conditions for the item, the item information may be a list of search results for the item, etc. Furthermore, the item information corresponding to the content of the viewing request may be any item information such as summary information, price information, or reviews. The server 10 simply reads the item information corresponding to the viewing request from the item information DB 103 and provides it to the terminal 20.

[0055] FIG. 7 shows an example of a display of item information and a recommendation list provided from the server 10 to the terminal 20. FIG. 7 shows, as an example, a case where a web page including item information 112 and a recommendation list display 113 is displayed in tab 1 of the web browser 110 by a web browser (tab browser) 110 installed on the terminal 20. FIG. 7 shows a web page resulting from, for example, a user accessing an item search site and performing an item search operation specifying search criteria. In this case, the web browser 110 displays item information 112, such as a list of search results for items in response to a browsing request based on the item search operation. In addition, a recommendation list is provided along with the item information 112. For example, when the item information 112 is scrolled downward, a recommendation list display 113 including multiple recommended items is displayed below the item information 112, as shown in FIG. 7.

[0056] The recommendation list display 113 shown in FIG. 7 displays the top three items in the recommendation ranking, i.e., items RI1, RI2, and RI3, as recommended items recommended to the user. Thus, in the display example shown in FIG. 7, three items in the recommendation list generated by the server 10 are displayed. For example, an overview of each item is displayed for items RI1, RI2, and RI3, and links to detailed information about each item. Also included is a scroll button 114 for scrolling through the items displayed in the recommendation list display 113. The recommendation list display 113 in FIG. 7 displays three items, but the recommendation list generated by the server 10 contains four or more items. Therefore, when the user operates the scroll button 114, the display range of the recommendation list display 113 moves in accordance with the order of the recommendation list (recommendation ranking), and the items are scrolled one by one. Also, an unnecessary selection button ds is displayed for each of items RI1, RI2, and RI3. When the unnecessary selection button ds is operated, for example, the item is deleted from the recommendation list display 113. When an item in the recommendation list display 113 is deleted by operating the unnecessary selection button ds, the recommendation list display 113 may be replenished with the next recommended item in the recommendation order.

[0057] The number of items displayed on the recommendation list display 113, i.e., the display range in which the items in the recommendation list transmitted from the server 10 are displayed at one time on the terminal 20, may be any number. Also, the display position of the recommendation list display 113 on the web page may be any position. That is, it is sufficient if the server 10 can provide the recommendation list in a format that allows the user to view it on the terminal 20.

[0058] In addition, in this embodiment, a portion of the recommendation list is displayed by scrolling in response to an operation on the scroll button 114 on the terminal 20, but the recommendation list may also be displayed by scrolling with a mouse or swipe, as a carousel display, or by dividing into multiple pages.

[0059] The terminal 20 may be a personal computer, a smartphone, a tablet, etc. The terminal 20 may display item information and recommendation lists in a dedicated application rather than the general-purpose web browser 110.

[0060] 4, the server 10 updates the user's recommendation history (step S14) by adding the items included in the recommendation list provided to the terminal 20 to the recommendation history of the corresponding user in the user information DB 104. Note that if the item viewing request is a request to view an item page and the item page is provided in response to the request, the user's viewing history may also be updated in response to the user viewing the item page.

[0061] Next, the server 10 determines whether or not a request to acquire an item page (item details) of a specific item has been received from the terminal 20 that has transmitted the recommendation list (step S15).

[0062] If the request to acquire an item page is received (step S15; Yes), the server 10 reads out the item page corresponding to the request from the item information DB 103 and transmits it to the terminal 20 (step S16).

[0063] As an example, when item RI3 (summary information of item 3) is selected on the screen shown in FIG. 7 , a request to acquire the item page of item 3 is sent from terminal 20 to server 10. Then, the item page of item 3 is sent from server 10 to terminal 20. In this case, for example, as shown in FIG. 8 , a new tab 2 opens in front of tab 1, and an item page including detailed information 115 of item 3 is displayed on tab 2. This item page serves as a response to the acquisition request. The detailed information 115 may include text showing details of item 3, price information, purchase information, reservation information, related information, word-of-mouth, review information, etc. If the item is a product or service, the detailed information 115 may be used to purchase or reserve the product or service. Note that the content of the detailed information 115 is arbitrary as long as it shows details of the corresponding item. Furthermore, when the detailed information tab is scrolled using the scroll bar or a mouse in the state shown in FIG. 8 , the information below can be viewed. By displaying such an item page on terminal 20, it can be inferred that the user is interested in the item (item 3 in this case), is concerned about it, is considering purchasing it, etc. Furthermore, when a link to an item page included in item information 112 is selected on the screen shown in FIG. 7, for example, when a specific item (item name, item summary, thumbnail, etc.) is selected from a list of search results for items, the corresponding item page is transmitted from server 10 and displayed on terminal 20. Note that the item page may be displayed on the same tab (tab 1 in FIG. 7, etc.) as the web page from which the link originates. Furthermore, a recommendation list may be displayed on the item page.

[0064] Then, the server 10 updates the user's browsing history by adding the item corresponding to the item page provided to the terminal 20 to the corresponding user's browsing history in the user information DB 104 and adding the recommendation ranking of the item from the recommendation list to the detailed browsing history (step S17).

[0065] If the request to acquire the item page has not been received (step S15; No), or after executing the process of step S17, the server 10 determines whether or not other operation information indicating that the user has performed another operation has been received from the terminal 20 (step S18). If other operation information has been received from the terminal 20 (step S18; Yes), the server 10 executes control according to the operation information (for example, control to update the web page) (step S19).

[0066] In step S19, for example, when operation information indicating that a user has purchased an item is received on an item page, the server 10 updates the user's purchase history by adding the purchased item to the corresponding user's purchase history in the user information DB 104, for example.

[0067] If no other operation information has been received from the terminal 20 (step S18; No), or after the processing of step S19, the server 10 returns to the processing of step S15, and repeats the processing according to the information received from the terminal 20. That is, the server 10 provides the recommendation list to the terminal 20 in step S13, and after causing the terminal 20 to present the recommendation list, waits for a new request from the terminal 20.

[0068] As described above, the recommendation system 100 and server 10 of this embodiment extract first-type items that satisfy a first condition from a plurality of items in response to an item viewing request from the terminal 20, and extract second-type items that satisfy a second condition different from the first condition from the plurality of items, and generate a recommendation list by sequentially arranging information on the extracted first-type items and information on the extracted second-type items that are not first-type items. The server 10 then transmits the generated recommendation list to the terminal 20, causing the terminal 20 to present the recommendation list. In this way, a recommendation list including first-type items extracted based on the first condition and second-type items extracted based on the second condition can be presented to the user of the terminal 20. Therefore, a variety of items can be presented while presenting useful items to the user, and items can be suitably recommended to the user.

[0069] Furthermore, the server 10 removes information about first-type items from the extracted second-type items, and then transmits information about second-type items that are not first-type items to the terminal 20, so no special operation or processing is required on the terminal 20 side.

[0070] (Embodiment 2) In the above embodiment, it has been described that the first condition, the second condition, the number N of items extracted based on the first condition, and the number n of items extracted based on the second condition are predetermined. However, this is not limiting, and it is also possible to analyze the access history to detailed information on items from a recommendation list provided to a user, output a report urging adjustment of at least one of the first condition and the second condition, or automatically adjust at least one of the first condition and the second condition. Note that N is included in the first condition, and n is included in the second condition.

[0071] For example, based on a history of accesses to detailed information about an item from a recommendation list presented in response to a past request, a first trend of attenuation of the remaining rate, which is the rate at which a user continues to glance at information about an item when the number of glanced items in the recommendation list presented on the terminal 20 is less than N, and a second trend of attenuation of the remaining rate when the number of glanced items is N or more, may be calculated. Then, it may be determined whether the second trend is significantly lower than the first trend based on a predetermined criterion. If the second trend is significantly lower, a report may be output prompting adjustment of at least one of the first and second conditions. Alternatively, if the second trend is significantly lower, at least one of the first and second conditions may be automatically adjusted. Adjusting the first and second conditions may mean, for example, relaxing or narrowing at least one of the first and second conditions, changing at least one of N and n, etc.

[0072] The first trend is the decay trend (decay rate, etc.) of the remaining rate when a user glances at a first-type item, and the second trend is the decay trend (decay rate, etc.) of the remaining rate when a user glances at a second-type item. The remaining rate tends to decay according to the number of items glanced at by the user. Furthermore, it is expected that the decay trend of the remaining rate will be stronger when the glanced item is a second-type item, which has a lower degree of personalization for the user, than when the glanced item is a first-type item. Therefore, the second trend is expected to be lower than the first trend. The second trend is compared with the first trend, and if the difference or the rate of decrease is higher than a predetermined criterion (e.g., a threshold), it is determined that the second trend is significantly lower than the first trend. In this case, it is expected that the user stopped viewing or glancing at the recommendation list because the items in the recommendation list switched from first-type items to second-type items. Therefore, a report is output prompting the user to adjust at least one of the first and second conditions. For example, a report may be output to encourage the user to increase the value of N included in the first condition, or the value of N may be automatically adjusted to a larger value. This allows more first-class items to be presented in the recommendation list, preventing a significant decrease in the remaining rate due to the number of items glanced at by the user. Furthermore, for example, the degree of user personalization of the second condition may be increased. This also prevents a decrease in the remaining rate due to the items in the recommendation list switching from first-class items to second-class items.

[0073] In addition, the predetermined judgment criteria (threshold value, etc.) for determining that the second trend is significantly lower than the first trend may be set, for example, based on a tendency to attenuate depending on the number of items glanced at by the user, regardless of the type of item.

[0074] Hereinafter, a second embodiment will be described in which the number N of items to be extracted based on the first condition is automatically adjusted based on the browsing history of the user's recommendation list. The configurations and processes of the recommendation system 100 and the server 10 of the second embodiment may be the same as those of the first embodiment, except for those described here.

[0075] 9 is a flowchart showing an example of an item extraction condition determination process executed by the server 10 of the second embodiment. The item extraction condition determination process is a process for determining whether or not the item extraction conditions used when generating a recommendation list in the server 10 are appropriate. The item extraction condition determination process is executed in response to the timing of reviewing the item extraction conditions or an operation by the administrator of the server 10. The item extraction condition determination process is executed by the processor 11 of the server 10 reading a program stored in the storage device 12.

[0076] In the item extraction condition determination process, the server 10 first acquires and tallys up the recommendation history of user information and detailed browsing history from the recommendation list of all users from the user information DB 104 (step S201).

[0077] Next, the server 10 calculates the remaining ratio for each number of glanced items in the recommendation list (step S202).

[0078] When detailed information about a specific item is accessed from the recommendation list, the detailed browsing history from the recommendation list includes information about the recommendation ranking of that item in the recommendation list. Accessing detailed information about an item from the recommendation list means that the user has at least browsed or glanced at the recommendation up to that item. Since the server 10 does not acquire the browsing status of the recommendation list provided to the terminal 20, it is not possible to identify up to which recommendation ranking the item was glanced. Therefore, in this embodiment, when detailed information about a specific item is accessed from the recommendation list, the remaining percentage is calculated assuming that the user has finished browsing the recommendation list at that point.

[0079] Specifically, in step S202, if it is determined from the detailed browsing history of the recommendation list that the user has accessed the detailed information of the t+1th item, this is considered to be the case where the user has browsed the detailed information of the t+1th item without accessing the detailed information of any of the items from the top to the tth item in the recommendation list. In this case, it is assumed that the user has glanced up to the tth item and remained. Then, the remaining number corresponding to the number of glanced items for all users is divided by the number of users to calculate the remaining ratio.

[0080] Next, the server 10 assumes that the remaining rate when the number of glanced items is less than N decays exponentially, and performs a regression analysis of the decay trend of the remaining rate when the number of glanced items is t less than N, to determine the coefficient a1 of the decay curve y=exp(-a1*t) as shown in Figure 10 (step S203).

[0081] Then, the server 10 extracts the remaining ratios when the number of glanced items is less than N, which are below the decay curve y=exp(-a1*t) derived in step S203 (step S204).

[0082] Furthermore, the server 10 performs a regression analysis on the attenuation tendency of the remaining ratio extracted in step S203, and obtains the coefficient a2 of the attenuation curve y=exp(-a2*t) as shown in FIG. 10 (step S205).

[0083] Then, the server 10 repeats the same process M times to obtain y=exp(-a m *t) coefficient a m is calculated (step S206).

[0084] In this way, it is possible to calculate the coefficients of the decay curve that extracts the low percentage of remaining percentages when the number of glanced items is less than N. Note that the value of M may be any value.

[0085] The a obtained in this way m is the first tendency of attenuation of the remaining rate when the number of glanced items is less than N.

[0086] Thereafter, the server 10 performs a regression analysis on the attenuation tendency of the remaining ratio when the number of glanced items is N or more, and obtains the coefficient b of the attenuation curve y=exp(-a1*t)*exp(-b*tN) as shown in Fig. 10 (step S207). The coefficient b obtained in this way is set as the second attenuation tendency of the remaining ratio when the number of glanced items is N or more.

[0087] Then, the server 10 calculates the coefficient b and the coefficient a m It is determined whether the second trend is significantly lower than the first trend based on the value of (step S208). For example, if the coefficient b is less than the coefficient a m If it is smaller than , the attenuation tendency of the remaining rate when the number of glanced items is N or more is lower than the attenuation tendency of the remaining rate extracted from the low rate of the remaining rate when the number of glanced items is less than N, so it is determined that the second tendency is significantly lower than the first tendency. In this case, the coefficient a m The P value of the regression analysis is (1 / 2) M+1 That is, the coefficient a m The condition may be that the probability is greater than a predetermined value.

[0088] Next, the server 10 outputs a report according to the determination result by displaying it on the display of the server 10 or by notifying the administrator contact person (step S209). This can prompt the administrator of the server 10 or the recommendation system 100 to adjust the item extraction conditions, i.e., adjust at least one of the first condition, the second condition, the value of N, and the value of M.

[0089] Thereafter, the server 10 determines whether to automatically adjust the item extraction conditions (step S210). Whether to automatically adjust the item extraction conditions may be set in advance by, for example, an administrator of the server 10 or the recommendation system 100.

[0090] If it is determined that the item extraction conditions will not be automatically adjusted (step S210; No), the item extraction condition determination process is terminated. If it is determined that the item extraction conditions will be automatically adjusted (step S210; Yes), the server 10 adjusts the item extraction conditions (first condition, second condition, at least one of the value of N and the value of M, etc.) based on the first trend and the second trend (step S211), and terminates the item extraction condition determination process. In step S211, if the determination result shows that the second trend is significantly lower than the first trend, the value of N can be adjusted to be increased.

[0091] As described above, the server 10 and the recommendation system 100 of the second embodiment output a report urging the user to adjust at least one of the first and second conditions based on the attenuation tendency of the remaining percentage in the recommendation list, or automatically adjust at least one of the first and second conditions, thereby making it possible to generate a recommendation list based on the remaining percentage in the user's past recommendation lists. That is, it becomes possible to provide a more suitable recommendation list based on the remaining percentage in the user's past recommendation lists. When automatically adjusting at least one of the first and second conditions, for example, the value N, which is the number of items extracted based on the first condition, can be automatically adjusted without the need for an administrator or the like to manually calculate or set it, thereby improving convenience.

[0092] (Variation) The present invention is not limited to the above-described embodiment, and various modifications and applications are possible. For example, some of the above-described embodiments may be omitted or replaced, or any configuration may be added.

[0093] In the above embodiment, the server 10 generates a recommendation list by excluding items that overlap with the first-type items from the second-type items. However, information about the first-type items and the second-type items may be transmitted from the server 10 to the terminal 20, and the terminal 20 may present a recommendation list by excluding items that overlap with the first-type items from the second-type items. For example, before presenting a second-type item on the terminal 20, it may be determined whether the item has already been presented as a first-type item, and if it has already been presented, the information about that item may be deleted and information about the second-type item may be presented. This allows the processing load on the server 10 to be distributed.

[0094] Furthermore, in the above embodiment, when it is determined that the user has accessed detailed information on the t+1th item in the recommendation list, it is determined that the user has glanced at the tth item from the top of the recommendation list, but the terminal 20 may monitor the browsing and glance status of the recommendation list and transmit the results to the server 10. In this way, the server 10 can identify the items glanced at by the user, and therefore can calculate the remaining ratio more accurately.

[0095] The server 10 and the terminal 20 can be realized using a normal computer, not a dedicated device. For example, the server 10 and the terminal 20 that execute the above-described processes may be configured by installing a program for executing the above-described functions on a computer from a recording medium that stores the program. Also, one server 10 may be configured by multiple computers operating in cooperation with each other.

[0096] Furthermore, the method for supplying the program to the computer is arbitrary. For example, the program may be supplied via a communication line, a communication network, a communication system, etc. For example, a non-transitory computer-readable recording medium storing the program for executing the above-described functions may be distributed.

[0097] Furthermore, if an OS (Operating System) provides some of the above functions, the functions other than those provided by the OS may be provided by a program.

[0098] The above-described embodiments are intended to facilitate understanding of the present invention and are not intended to limit the present invention. The flowcharts, sequences, elements included in the embodiments, and their arrangements, materials, conditions, shapes, sizes, etc., described in the embodiments are not limited to those illustrated and can be modified as appropriate. Furthermore, configurations shown in different embodiments can be partially substituted or combined with each other.

[0099] (Invention realized by this embodiment) As described above, this embodiment realizes the following inventions.

[0100] (Appendix 1) A server that recommends items to a user, extracting, in response to a request from the terminal, a first type item that satisfies a first condition from the plurality of items, and extracting, from the plurality of items, a second type item that satisfies a second condition different from the first condition; The information on the extracted first-type items and the information on the extracted second-type items that are not the first-type items are arranged in order and presented to the terminal, and a new request from the terminal is awaited. server.

[0101] (Appendix 2) The extracted information on the second type item is transmitted from the server to the terminal; The terminal deletes the items already presented, and then presents information about the second-type items that are not the first-type items. The server described in Appendix 1.

[0102] (Appendix 3) The server removes information about the first-type items from the extracted second-type items, and then transmits information about the second-type items that are not the first-type items to the terminal. The server described in Appendix 1.

[0103] (Appendix 4) As the first condition, a condition that the degree of match for the user who uses the terminal is the top N is set, Based on a history of accesses to detailed information of an item from information of the item presented in response to a past request, a first tendency of attenuation of a remaining rate, which is the rate at which a user continues to glance at information of an item when the number of glanced items glanced by the user in the information of the item presented on the terminal is less than N, and a second tendency of attenuation of the remaining rate when the number of glanced items is equal to or greater than N; determining whether the second trend is significantly lower than the first trend based on a predetermined criterion; If the difference is significantly lower, a report is output to prompt adjustment of at least one of the first condition and the second condition. 2. A server as set forth in Appendix 1, 2 or 3.

[0104] (Appendix 5) When detailed information on the t+1th item in the item information presented in response to a request from the terminal is accessed, it is determined that the user has continued to glance up to the information on the tth item. A server as described in Appendix 4.

[0105] (Appendix 6) If the second tendency is significantly lower than the first tendency, at least one of the first condition and the second condition is adjusted. 1. A server as described in Appendix 4 or 5.

[0106] (Appendix 7) The first condition is a condition that is more personalized for the user who uses the terminal than the second condition. 1. A server according to any one of appendices 1 to 6.

[0107] (Appendix 8) A recommendation method in which a server recommends an item to a user, comprising: extracting, in response to a request from the terminal, a first type item that satisfies a first condition from the plurality of items, and extracting, from the plurality of items, a second type item that satisfies a second condition different from the first condition; The information on the extracted first-type items and the information on the extracted second-type items that are not the first-type items are arranged in order and presented to the terminal, and a new request from the terminal is awaited. Recommendation method.

[0108] (Appendix 9) Computer, extracting, in response to a request from the terminal, a first type item that satisfies a first condition from the plurality of items, and extracting, from the plurality of items, a second type item that satisfies a second condition different from the first condition; The information on the extracted first-type items and the information on the extracted second-type items that are not the first-type items are arranged in order and presented to the terminal, and a new request from the terminal is awaited. A program that acts as a server. [Industrial Applicability]

[0109] According to the present invention, it is possible to provide a server, a recommendation method, and a program that can recommend items to users in an appropriate manner. [Explanation of symbols]

[0110] 10 Servers 20 terminals 11, 21 processors 12, 22 Storage device 13, 23 Communication Interface 14, 24 input devices 15, 25 output devices 100 Recommendation System 101 Transmitter / Receiver 102 Recommendation list generation unit 1021 Type 1 Item Extraction Unit 1022 Type 2 Item Extraction Unit 103 Item Information DB 104 User Information DB 110 Web Browser 112 Item Information 113 Recommendation list display 114 Scroll button 115 More Information RI1, RI2, RI3 items (recommended items) ds Unnecessary selection button

Claims

1. A server that recommends items to a user, extracting, in response to a request from the terminal, a first type item that satisfies a first condition from the plurality of items, and extracting, from the plurality of items, a second type item that satisfies a second condition different from the first condition; The information on the extracted first-type items and the information on the extracted second-type items that are not the first-type items are arranged in order and presented to the terminal, and a new request from the terminal is awaited. server.

2. The information of the extracted second-type items is transmitted from the server to the terminal; After deleting the items already presented on the terminal side, information on the second-type items that are not the first-type items is presented. The server of claim 1 .

3. The server removes information about the first-type items from the extracted second-type items, and then transmits information about the second-type items that are not the first-type items to the terminal. The server of claim 1 .

4. As the first condition, a condition that the degree of match for the user who uses the terminal is the top N is set, Based on a history of accesses to detailed information of an item from information of the item presented in response to a past request, a first tendency of attenuation of a remaining rate, which is the rate at which a user continues to glance at information of an item when the number of glanced items glanced by the user in the information of the item presented on the terminal is less than N, and a second tendency of attenuation of the remaining rate when the number of glanced items is equal to or greater than N, determining whether the second trend is significantly lower than the first trend based on a predetermined criterion; If the difference is significantly lower, a report is output to prompt adjustment of at least one of the first condition and the second condition. The server of claim 1 .

5. When detailed information on the t+1th item in the item information presented in response to a request from the terminal is accessed, it is determined that the user has continued to glance up to the tth item information. The server of claim 4.

6. If the second tendency is significantly lower than the first tendency, at least one of the first condition and the second condition is adjusted. The server of claim 4.

7. The first condition is a condition that is more personalized for the user who uses the terminal than the second condition. The server according to any one of claims 1 to 6.

8. A recommendation method in which a server recommends an item to a user, comprising: extracting, in response to a request from the terminal, a first type item that satisfies a first condition from the plurality of items, and extracting, from the plurality of items, a second type item that satisfies a second condition different from the first condition; The information on the extracted first-type items and the information on the extracted second-type items that are not the first-type items are arranged in order and presented to the terminal, and a new request from the terminal is awaited. Recommendation method.

9. Computer, extracting, in response to a request from the terminal, a first type item that satisfies a first condition from the plurality of items, and extracting, from the plurality of items, a second type item that satisfies a second condition different from the first condition; The information on the extracted first-type items and the information on the extracted second-type items that are not the first-type items are arranged in order and presented to the terminal, and a new request from the terminal is awaited. A program that acts as a server.

Images