Recommendation determination device, display information recommendation method, and program

Abstract

To provide a recommendation determination device, a recommendation method of display information, and a program capable of providing new information.SOLUTION: According to an embodiment, a recommendation determination device includes a communication unit and a processor. The communication unit communicates with a display device. The processor acquires information on preference of a target person specified by identification information acquired from a person who can be visually recognized by the display device, and supplies, to the display device, a display instruction of one piece of display information determined from display information to be a display candidate retrieved on the basis of recommended content extracted from preference content of a reference user having preference different from the target person.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] An embodiment of the present invention relates to a recommendation determination device, a display information recommendation method, and a program. [Background technology]

[0002] Conventionally, there are systems that display advertisements to users at devices such as automatic ticket gates used in transportation such as railways. For example, automatic ticket gates display advertisements related to ticket information presented by users. Furthermore, general advertisement display systems are configured to display advertisements that match the preferences of users.

[0003] However, since a user's preferences are estimated based on pre-registration information or past behavioral history, advertisements that match such preferences tend to be similar. As a result, if the same types of advertisements are repeatedly displayed on devices that the user frequently uses, the user may become bored with the displayed advertisements or lose interest in the displayed content. Furthermore, there is a problem that advertisement providers cannot effectively provide advertisements that interest users, and cannot present useful information that the user has not yet recognized. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 2023-163494 Summary of the Invention [Problem to be solved by the invention]

[0005] In order to solve the above problems, an object is to provide a recommendation determination device, a display information recommendation method, and a program that can provide new information. [Means for solving the problem]

[0006] According to an embodiment, a recommendation determination device includes a communication unit and a processor. The communication unit communicates with a display device. The processor acquires information about the preferences of a target person identified by identification information acquired from a person who can view the display device, and provides the display device with a display instruction for one piece of display information determined from display information that is a display candidate searched based on recommended content extracted from preference content of a reference user who has different preferences from the target person. [Brief explanation of the drawings]

[0007] [Figure 1] FIG. 1 is a diagram illustrating an example of the configuration of an information recommendation system according to an embodiment. [Figure 2] FIG. 2 is a block diagram showing an example of the configuration of a recommendation determination server in the information recommendation system according to the embodiment. [Figure 3] FIG. 3 is a block diagram showing an example of the configuration of a recommended information server in the information recommendation system according to the embodiment. [Figure 4] FIG. 4 is a block diagram showing an example of the configuration of an automatic ticket gate in the information recommendation system according to the embodiment. [Figure 5] FIG. 5 is a sequence diagram for explaining the flow of advertisement provision processing (information recommendation processing) in the entire information recommendation system according to the embodiment. [Figure 6] FIG. 6 is a flowchart illustrating an example of a process of updating a user database by a recommended information server in the information recommendation system according to the embodiment. [Figure 7] FIG. 7 is a diagram showing an example of movement history information based on log data stored by the recommended information server in the information recommendation system according to the embodiment. [Figure 8] FIG. 8 is a diagram showing an example of a preference information pool for each user stored by the recommended information server in the information recommendation system according to the embodiment. [Figure 9] FIG. 9 is a diagram showing an example of a preference vector for each user stored by the recommended information server in the information recommendation system according to the embodiment. [Figure 10]FIG. 10 is a flowchart illustrating an example of a process in which the recommended information server in the information recommendation system according to the embodiment instructs the automatic ticket gate to display an advertisement. [Figure 11] FIG. 11 is a flowchart illustrating an example of a process for selecting a referring user by the recommendation determination server in the information recommendation system according to the embodiment. [Figure 12] FIG. 12 is a diagram showing an example in which the recommendation determination server in the information recommendation system according to the embodiment focuses on the first feature amount and adds other users similar to the target user to the reference user list. [Figure 13] Figure 13 is a diagram showing an example in which other users who have some common preferences with the subject calculated from features other than the first feature but whose preference vectors have low similarity to the subject are left in the reference user list shown in Figure 12. [Figure 14] FIG. 14 is a diagram showing an example in which the recommendation determination server in the information recommendation system according to the embodiment focuses on the second feature amount and adds other users similar to the target user to the reference user list. [Figure 15] FIG. 15 is a diagram showing an example in which other users whose preference vectors have low similarity to the subject calculated from feature amounts other than the second feature amount are left in the reference user list shown in FIG. [Figure 16] FIG. 16 is a flowchart illustrating an example of a content ranking process performed by the recommendation determination server in the information recommendation system according to the embodiment. [Figure 17] FIG. 17 is a diagram showing an example of a target user's preference information pool (a group of preference contents) and other users' preference information pools (a group of preference contents). [Figure 18] FIG. 18 is a diagram showing an example of a list of contents extracted from the preference information pool of the target person shown in FIG. 17 and the preference information pool of other users. [Figure 19] FIG. 19 is a diagram showing an example of the feature amount of the subject and the feature amount of each piece of content shown in FIG. [Figure 20]FIG. 20 is a diagram showing an example in which each content is ranked based on the feature amount of the subject shown in FIG. 19 and the feature amount of each content. [Figure 21] FIG. 21 is a flowchart illustrating an example of an advertisement determination process by the recommendation determination server in the information recommendation system according to the embodiment. [Figure 22] FIG. 22 is a diagram showing an example of keywords set for each advertisement managed by the advertisement DB server in the information recommendation system according to the embodiment. [Figure 23] FIG. 23 is a diagram showing examples of advertisements that are candidates for display selected from the recommended content shown in FIG. 20 and the keywords of each advertisement shown in FIG. [Figure 24] FIG. 24 is a diagram showing an example of an advertisement displayed on a display of an automatic ticket gate in the information recommendation system according to the embodiment. [Figure 25] FIG. 25 is a flowchart illustrating an example of a user database update process performed by the recommendation determination server in the information recommendation system according to the embodiment. [Figure 26] FIG. 26 is a flowchart illustrating an example of a user database update process performed by the recommendation determination server in the information recommendation system according to the embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0008] Hereinafter, embodiments will be described with reference to the drawings. FIG. 1 is a diagram illustrating an example of the configuration of an information recommendation system according to an embodiment. The information recommendation system 1 according to this embodiment is a system that provides advertisements to users (users) who use transportation such as railways. Based on information such as the user's preference information and behavioral history, the information recommendation system 1 recommends advertisements that will lead to new discoveries for the user as advertisements to be displayed.

[0009] Furthermore, the information recommendation system 1 extracts preference information of other users who share some of the same preferences as the user (consumer) who is the target of recommendation, and recommends advertisements of content different from the preferences of the target of recommendation from the preference information of the other users. As a specific example of operation, the information recommendation system 1 illustrated in Fig. 1 selects advertisements to be recommended to users passing through a ticket gate where an automatic ticket gate 11 is installed, and displays the recommended advertisements on a display device (for example, a display on the automatic ticket gate) that is visible to users passing through the ticket gate.

[0010] However, the information recommendation system 1 is not limited to the configuration shown in Fig. 1. For example, instead of an automatic ticket gate, the information recommendation system 1 may be configured using a device that reads the user's identification information and connects to a display device that is visible to the user.

[0011] 1, the information recommendation system 1 is configured with an automatic ticket gate 11, a recommendation determination server (recommendation determination device) 12, a recommendation information server 13, and an advertisement DB server 14. The recommendation determination server 12 and the recommendation information server 13 may be configured as a single server device. Also, the recommendation determination server 12 or the recommendation information server 13 and the advertisement DB server 14 may be configured as a single server device.

[0012] The automatic ticket gate 11 is an example of a device equipped with a display device 11a as a display device visible to a user. The automatic ticket gate 11 is installed, for example, at a ticket gate in a station. The automatic ticket gate 11 reads an ID presented (carried) by a user and performs a passage control process (ticket gate process) that controls the passage (entrance and exit) of the user identified by the ID. For example, the automatic ticket gate 11 performs the ticket gate process for the user and displays an advertisement on the display device 11a according to the user passing through (entering or exiting).

[0013] The display 11a may be installed at a position separate from the automatic ticket gate 11. In other words, the display 11a may be a display device that is located at a position visible to users passing through the passage formed by the automatic ticket gate 11. For example, the display 11a may be a display device that is placed in the direction of travel of users passing through the passage formed by the automatic ticket gate 11. When the display 11a is installed at a position separate from the housing of the automatic ticket gate 11, the display 11a and the automatic ticket gate 11 may be configured to be connected for communication wirelessly or via a wire.

[0014] In the configuration example shown in Fig. 1, the automatic ticket gate 11 is communicably connected to the recommendation determination server 12, the recommendation information server 13, and the advertisement DB server 14. The automatic ticket gate 11 also has a display 11a that is arranged so that it can be seen by users passing through. The automatic ticket gate 11 is also configured to read the ID indicated by the boarding tool M that the user presents (carries) as a boarding ticket. The boarding tool M may be anything that indicates the user's ID, such as an IC card, a mobile terminal, a credit card, or a ticket with a two-dimensional code printed on it (a code ticket).

[0015] The recommendation determination server 12 includes a computer (information processing device) that executes various processes using programs. The recommendation determination server 12 communicates with the automatic ticket gate 11, which acquires identification information (ID) that uniquely identifies a user. The recommendation determination server 12 also communicates with the recommendation information server 13 and the advertisement DB server 14. The recommendation determination server 12 has a function of determining the advertisement to be presented to the user of the ID acquired by the automatic ticket gate 11.

[0016] For example, the recommendation determination server 12 acquires information about a user (target person) passing through a ticket gate where the automatic ticket gate 11 is installed and information about users other than the target person from the recommendation information server 13. The recommendation determination server 12 selects (recommends) content (keywords, etc.) that will stimulate new discoveries for the target user based on the information acquired from the recommendation information server 13. The recommendation determination server 12 acquires information about advertisements corresponding to content to be recommended to the user from the advertisement DB server 14, and determines advertisements to be displayed on the automatic ticket gate 11 based on the information acquired from the advertisement DB server 14.

[0017] The recommendation information server 13 includes a computer (information processing device) that executes various processes using programs. The recommendation information server 13 is a server device configured to communicate with the automatic ticket gate 11 and the recommendation determination server 12. The recommendation information server 13 manages information used by the recommendation determination server 12 to select content to recommend to individual users. The recommendation information server 13 acquires user information (log data) associated with IDs from the automatic ticket gate 11. For example, the recommendation information server 13 associates the user ID acquired from the automatic ticket gate 11 with the user's entry and exit information (movement information) and stores the associated information.

[0018] Furthermore, the recommendation information server 13 estimates preference information indicating the preferences of the user based on the user's entry / exit information (movement information) and the like. After estimating the user's preferences, the recommendation information server 13 stores the preference information for each user. Furthermore, the recommendation information server 13 creates and stores a preference vector indicating the characteristics of the preferences of each user.

[0019] The advertisement DB server 14 includes a computer (information processing device) that executes various processes using programs. The advertisement DB server 14 is a server device configured to communicate with the automatic ticket gate 11 and the recommendation determination server 12. The advertisement DB server 14 manages information about a plurality of advertisements that can be displayed on the display 11a of the automatic ticket gate 11. The advertisement DB server 14 has a database (DB) that stores keywords and the like that indicate content related to each advertisement.

[0020] For example, the advertisement DB server 14 has a function of extracting a group of advertisements corresponding to a specified keyword (or content), and supplies advertisement data indicating a group of advertisements corresponding to the keyword specified by the recommendation determination server 12 to the recommendation determination server 12. The advertisement DB server 14 also has a function of supplying display data of an advertisement specified by the automatic ticket gate 11 equipped with the display 11a to the automatic ticket gate 11.

[0021] Next, the configuration of the recommendation determination server 12 in the information recommendation system 1 according to the embodiment will be described. FIG. 2 is a block diagram showing an example of the configuration of the recommendation determination server 12 in the information recommendation system 1 according to the embodiment. As shown in FIG. 2, the recommendation determination server 12 includes a processor 21, a ROM 22, a RAM 23, a storage unit 24, a communication interface (I / F) 25, and the like. The processor 21 controls the recommendation determination server 12 and performs data processing, etc. The processor 21 realizes various functions by executing programs. The processor 21 is, for example, a CPU. The processor 21 is configured to execute programs stored in the ROM 22 or the storage unit 24 using the RAM 23.

[0022] The ROM 22 is a non-volatile memory that stores control programs and control data that govern basic operations. The RAM 23 is a volatile memory that functions as a working memory that temporarily stores working data.

[0023] The storage unit 24 is a memory that stores data. The storage unit 24 is configured as a rewritable nonvolatile storage device. The storage unit 24 is configured as, for example, a hard disk drive (HDD) or a solid state drive (SSD).

[0024] The communication interface 25 is a communication interface for performing data communication. For example, the communication interface 25 is a network interface that communicates with each device via a network. In this embodiment, the communication interface 25 has interfaces for communicating with the automatic ticket gate 11, the recommendation information server 13, and the advertisement DB server 14.

[0025] The processor 21 includes a user search unit 21a, a ranking unit 21b, and an advertisement determination unit 21c as examples of functions realized by executing a program. The user search unit 21a includes a function of referring to multiple feature amounts indicating the preferences of a recommendation target person, searching for users with similar feature amounts, and selecting users from the searched users whose other preference feature amounts differ from the preference feature amounts of the recommendation target person. The ranking unit 21b is a function of assigning an order (rank) to information (content) that is as far away as possible from the preferences of the recommendation target person from the preference information of other selected users. The advertisement determination unit 21c is a function of determining an advertisement to display based on the recommendation ranking for the information and the advertisement display history for the recommendation target person.

[0026] Next, the configuration of the recommended information server 13 in the information recommendation system according to the embodiment will be described. FIG. 3 is a block diagram showing an example of the configuration of the recommended information server 13 in the information recommendation system according to the embodiment. As shown in FIG. 3, the recommendation information server 13 includes a processor 31, a ROM 32, a RAM 33, a storage unit 34, a communication interface (I / F) 35, and the like. The processor 31 performs information management, data processing, and the like in the recommendation information server 13. The processor 31 realizes various functions by executing programs. The processor 31 is, for example, a CPU. The processor 31 includes, for example, a feature calculation unit 31a as a function realized by executing a program.

[0027] The ROM 32 is a non-volatile memory that stores control programs and control data that govern basic operations. The RAM 33 is a volatile memory that functions as a working memory that temporarily stores working data.

[0028] The storage unit 34 is a memory that stores data. The storage unit 34 is configured as a rewritable nonvolatile storage device. The storage unit 34 is configured as, for example, a hard disk drive (HDD) or a solid state drive (SSD).

[0029] The storage unit 34 has a content DB 34a and a user DB 34b. The content DB 34a stores data related to content corresponding to keywords for selecting advertisements. The user DB 34b checks information for each user (user information) in association with the user ID.

[0030] The communication interface 35 is a communication interface for performing data communication. For example, the communication interface 35 is a network interface that communicates with each device via a network. In this embodiment, the communication interface 35 has an interface for communicating with the automatic ticket gate 11 and the recommendation determination server 12.

[0031] The processor 31 of the recommendation information server 13 includes a feature calculation unit 31a as an example of a function realized by executing a program. The feature calculation unit 31a calculates the feature of a user or content. For example, the feature calculation unit 31a calculates a vector (preference vector) as a feature indicating the user's preferences, and stores the calculated preference vector in the storage unit 34 in association with the user.

[0032] Next, the configuration of the automatic ticket gate 11 in the information recommendation system 1 according to the embodiment will be described. FIG. 4 is a block diagram showing an example of the configuration of the automatic ticket gate 11 in the information recommendation system 1 according to the embodiment. In the configuration example shown in FIG. 4, the automatic ticket gate 11 includes a control unit 41, a data memory 42, a communication unit 43, an ID reader 44, a passage control unit 45, and a display 11a.

[0033] The control unit 41 is responsible for overall control of the automatic ticket gate 11. The control unit 41 has a processor, memory, various interfaces, etc. The processor is a CPU that executes programs, etc. The memory includes RAM, ROM, and rewritable non-volatile memory (e.g., EEPROM, FROM, etc.). The control unit 41 realizes various functions by having the processor execute programs stored in the memory. For example, the control unit 41 is connected to each part within the automatic ticket gate 11 via an interface and controls each part within the automatic ticket gate 11.

[0034] The data memory 42 stores various types of data. For example, the data memory 42 may store a program for executing ticket inspection processing or display processing. The data memory 42 may also store fare data and other fee data.

[0035] The communication unit 43 is a communication unit for communicating with each device in the system. The communication unit 43 is, for example, a network interface. According to the configuration example shown in Fig. 1, the communication unit 43 includes interfaces for communicating with the recommendation determination server 12, the recommendation information server 13, and the advertisement DB server 14.

[0036] The ID reader 44 is configured as a device that reads the ID of a user. The ID reader 44 acquires information including the ID from the riding tool M presented (held) by the user and outputs the acquired information including the ID to the control unit 41. For example, if the riding tool M presented by the user is a code ticket with a printed code image, the ID reader 44 is configured as a device that reads the ID from the code image on the code ticket presented by the user. If the riding tool M is an IC card, the ID reader 44 is configured as a card reader / writer that reads information including the ID from the IC card. If the riding tool M is a mobile terminal, the ID reader 44 is configured as a reader / writer that reads information including the ID from the mobile terminal.

[0037] The passage control unit 45 controls the passage of users through the ticket gate (passageway) where the automatic ticket gate 11 is installed. The passage control unit 45 includes a drive mechanism for opening and closing doors to block passage through the passageway. When the passage control unit 45 permits a user to pass, it opens the door in accordance with instructions from the control unit 41 to encourage the user to pass, and when the passage control unit 45 prohibits the user from passing, it closes the door in accordance with instructions from the control unit 41 to prohibit the user from passing.

[0038] The display device 11a is a display device that displays various kinds of information, including advertisements, to users. The display device 11a is provided in a position visible to users passing through the ticket gate (passageway) where the automatic ticket gate 11 is installed. The display device 11a is provided, for example, on the housing of the automatic ticket gate 11 that forms the passageway. However, the display device 11a may also be a display device that is provided in a location separate from the housing of the automatic ticket gate 11.

[0039] The display device 11a displays an image of an advertisement or the like on a display screen in accordance with a control instruction from the control unit 41. For example, the display device 11a displays an image based on display data of the advertisement acquired from the advertisement DB server 14, thereby displaying the advertisement.

[0040] Next, a process of recommending an advertisement (information recommendation process) by the information recommendation system 1 according to the embodiment will be described.

[0041] FIG. 5 is a sequence diagram for schematically explaining the flow of a process (information recommendation process) for recommending an advertisement by the information recommendation system 1 according to the embodiment. Here, we will explain the operation of displaying an advertisement for a user on the display 11a of the automatic ticket gate 11 when the user passes through a ticket gate at a station. First, the user presents the boarding tool M to the automatic ticket gate 11 in order to pass through the ticket gate.

[0042] When the control unit 41 of the automatic ticket gate 11 detects a user attempting to pass through the ticket gate with a sensor, the control unit 41 acquires information including the user's ID from the boarding tool M carried by the user with the ID reader 44 (step ST11). For example, the control unit 41 reads the user's ID and information required for ticket gate processing from the boarding tool M.

[0043] When the boarding tool M reads the information including the ID, the control unit 41 of the automatic ticket gate 11 performs ticket gate processing for the user (step ST12). In the ticket gate processing, the control unit 41 determines whether or not the user is permitted to pass based on the information read from the boarding tool M, and controls the user's passage depending on whether or not the user is permitted to pass. For example, the control unit 41 may determine whether or not the user is permitted to pass based on the information read from the boarding tool M, or may request an external server to determine whether or not the user is permitted to pass based on the ID.

[0044] If the control unit 41 of the automatic ticket gate 11 determines that the user should be permitted to pass, it opens the door using the passage control unit 45 to allow the user to pass, and if it determines that the user should not be permitted to pass, it closes the door using the passage control unit 45 to prevent the user from passing. Here, in order to explain the operation of displaying advertisements to users passing through the ticket gate, the following operation will be explained assuming that the user is permitted to pass in the ticket gate processing.

[0045] When the control unit 41 of the automatic ticket gate 11 performs ticket gate processing for the user, the control unit 41 transmits information (log data) indicating the result of the ticket gate processing to the recommendation information server 13 via the communication unit 43 (step ST13). Here, the log data indicating the result of the ticket gate processing includes the user's ID and entry / exit information (movement information) indicating entry or exit at the ticket gate.

[0046] The recommendation information server 13 receives log data from the automatic ticket gate 11 via the communication interface 35. When the processor 31 of the recommendation information server 13 receives the log data from the automatic ticket gate 11, it updates the user database (DB) 34b based on the received log data (step ST14). The user database 34b stores user data including a preference information pool indicating the preferences of each user and a preference vector indicating the feature quantities of the preferences. An example of the update process of the user database 34b will be described in detail later.

[0047] Furthermore, when the control unit 41 of the automatic ticket gate 11 acquires the user's ID, it transmits the user's ID to the recommendation determination server 12 via the communication unit 43 in order to acquire advertisements targeted to the user (step ST15). The control unit 41 may transmit the user's ID when it determines that the user is permitted to pass through the ticket gate process, or may transmit the user's ID immediately after acquiring the ID without waiting for the ticket gate process to be completed.

[0048] The recommendation determination server 12 receives the user ID from the automatic ticket gate 11 via the communication interface 25. When the processor 21 of the recommendation determination server 12 receives the ID from the automatic ticket gate 11, it communicates with the recommendation information server 13 via the communication interface 25 and requests information on the user (target person) identified by the ID received from the automatic ticket gate 11 from the recommendation information server 13 (step ST16).

[0049] For example, the processor 21 of the recommendation determination server 12 requests, as user information, user data and advertisement display history stored in the user database 34b by the recommendation information server 13. The user data requested from the recommendation information server 13 is information that includes a preference vector indicating the feature amounts obtained by breaking down the preferences of the user into multiple types of feature amounts.

[0050] The recommendation information server 13 receives the request for user data from the recommendation determination server 12 via the communication interface 35. The processor 31 of the recommendation information server 13 extracts user information (user data and advertisement display history) corresponding to the received requested ID from the user database 34b, and transmits the extracted user information to the recommendation determination server 12 via the communication interface 35 (step ST17).

[0051] The recommendation determination server 12 receives user information including user data of the user from the recommendation information server 13 via the communication interface 25. The processor 21 of the recommendation determination server 12 stores the user data acquired from the recommendation information server 13 in a memory such as the RAM 23. The processor 21 of the recommendation determination server 12 performs a process of selecting a user to refer to in order to determine an advertisement to be provided to the user from the user data included in the user information acquired from the recommendation information server 13 (steps ST21-25).

[0052] The processor 21 of the recommendation determination server 12 selects one feature from multiple types of feature represented by multiple preference vectors in the user data (step ST21). After selecting one feature, the processor 21 requests the recommendation information server 13 for user data (preference vectors) of multiple other users selected using the selected feature as a condition (step ST22). For example, the processor 21 requests the recommendation information server 13 for user data of multiple other users having feature similar to the feature of the selected user.

[0053] The recommendation information server 13 receives a request for user data of other users from the recommendation determination server 12 via the communication interface 35. The processor 31 of the recommendation information server 13 searches the user database 34b for user data of multiple (a predetermined number) other users based on the conditions specified in the request from the recommendation determination server 12. For example, the processor 31 selects, from the user database 34b, user data of a predetermined number of other users that are similar in one feature amount that is the specified condition. The processor 31 reads the user data of the predetermined number of other users selected based on the specified condition from the user database 34b and transmits it to the recommendation determination server 12 (step ST23).

[0054] The recommendation determination server 12 receives, via the communication interface 25, the user data of multiple other users searched based on the conditions specified by the recommendation information server 13. The processor 21 of the recommendation determination server 12 uses a preference vector indicating the preferences of the other users contained in the user data of the multiple other users acquired from the recommendation information server 13 and a usage vector indicating the user's own preferences to determine the degree of deviation between the user's preferences and those of the other users (step ST24). The processor 21 of the recommendation determination server 12 performs processing to update (or generate) a list of users to be referenced in order to determine an advertisement to be provided (recommended) to the user, based on the degree of deviation between the user's preferences and those of the other users (step ST25).

[0055] Here, "deviation" is an evaluation value that represents the degree of dissimilarity, and "similarity" is an evaluation value that represents the degree of similarity. In other words, "deviation" and "similarity" are evaluation values ​​that are inversely correlated. For example, "high deviation" indicates "low similarity," and "low deviation" indicates "high similarity."

[0056] The processing of ST21-25 by the recommendation determination server 12 and the recommendation information server 13 is repeatedly executed for each feature (or a predetermined number of feature) in the user data of the user. The processor 21 of the recommendation determination server 12 repeatedly executes the processing of steps ST21-25 for each of the multiple feature, thereby generating a list of users (reference user list) to be referenced in selecting advertisements to be provided to the user.

[0057] After generating the reference user list, the processor 21 of the recommendation determination server 12 requests the recommendation information server 13 to send a preference information group (preference information pool) listing the preferences of each of the other users listed in the reference user list (step ST26).

[0058] The recommendation information server 13 receives a request for the preference information pool of each user on the reference user list from the recommendation determination server 12 via the communication interface 35. The processor 31 of the recommendation information server 13 reads out the preference information pool of each user specified in the request from the recommendation determination server 12 from the user database 34b. The processor 31 of the recommendation information server 13 transmits the preference information pool read out from the user database 34b to the recommendation determination server 12 (step ST27).

[0059] The recommendation determination server 12 receives the preference information pool from the recommendation information server 13 via the communication interface 25. The processor 21 of the recommendation determination server 12 calculates the degree of deviation from the user's preferences for each piece of information (content) included in the preference information pool acquired from the recommendation information server 13 (steps ST28-29).

[0060] The processor 21 of the recommendation determination server 12 selects one piece of information from each piece of information (content) included in the preference information pool acquired from the recommended information server 13 (step ST28). Here, in order to rank content that deviates from the user's preferences, information that is not included in the preference information pool of the user (i.e., the user's preferences) is selected. That is, the processor 21 selects one piece of information from the content included in the preference information pool acquired from the recommended information server 13 that is not included in the preference information pool of the user (i.e., the user's preferences).

[0061] When processor 21 of recommendation determination server 12 selects one piece of content, it calculates the degree of discrepancy between the selected content and the preferences of the user (target person) (step ST29). Processor 21 calculates the degree of discrepancy based on the feature amounts of the user's preferences and the feature amounts of the selected content. As a specific example, processor 21 calculates the degree of discrepancy based on the cosine similarity (COS similarity) between a preference vector indicating the feature amounts of the user's preferences and a vector indicating the feature amounts of the selected content.

[0062] The processor 21 of the recommendation determination server 12 repeatedly executes the processes of steps ST28 and ST29 to calculate the degree of deviation from the user's preferences for a plurality of pieces of content.

[0063] After calculating the deviation degrees for the plurality of pieces of content, the processor 21 of the recommendation determination server 12 ranks each piece of content based on the calculated deviation degrees (step ST31). For example, the processor 21 ranks each piece of content by creating a ranking of the pieces of content in descending order of deviation degrees.

[0064] After ranking the content, the processor 21 of the recommendation determination server 12 requests data of advertisements (data indicating the details of advertisements) related to the content for each of the top-ranked content from the advertisement DB server 14 (step ST32). For example, the processor 21 requests data of a predetermined number of advertisements for each of the top-ranked content from the advertisement DB server 14.

[0065] The advertisement DB server 14 receives a request for advertisement data related to the content from the recommendation determination server 12 via the communication interface. In response to the request from the recommendation determination server 12, the advertisement DB server 14 extracts advertisement data for a predetermined number of specified contents from the advertisement database, and transmits the extracted advertisement data for the predetermined number of contents to the recommendation determination server 12 (step ST33).

[0066] The recommendation determination server 12 receives advertisement data from the advertisement DB server 14 via the communication interface 25. When the processor 21 of the recommendation determination server 12 acquires a predetermined number of advertisement data for each content from the advertisement DB server 14, it determines advertisements (display advertisements) to be provided to the user from the advertisement data (step ST34).

[0067] For example, the processor 21 of the recommendation determination server 12 lists advertisement data in order according to the ranking of the content. From the advertisement data listed in order, the processor 21 determines the advertisement with the smallest display history in the advertisement display history included in the user data of the user as the advertisement to be displayed. For example, if there is an advertisement with zero advertisement display history, the processor 21 determines the advertisement at the top of the list of advertisements with zero advertisement display history as the advertisement to be displayed. When the processor 21 of the recommendation determination server 12 determines the advertisement to be displayed, the processor 21 transmits a display instruction for the advertisement to the automatic ticket gate 11 (step ST35).

[0068] The automatic ticket gate 11 receives the designation of the display advertisement from the recommendation determination server 12 via the communication unit 43. When the control unit 41 of the automatic ticket gate 11 receives the designation of the display advertisement from the recommendation determination server 12, it requests the display data of the designated display advertisement from the advertisement DB server 14 (step ST36).

[0069] When the advertisement DB server 14 receives a request for advertisement display data from the automatic ticket gate 11, it extracts the display data of the specified advertisement from the advertisement database and transmits it to the automatic ticket gate 11 (step ST37).

[0070] The automatic ticket gate 11 receives the advertisement display data from the advertisement DB server 14 via the communication unit 43. When the control unit 41 of the automatic ticket gate 11 receives the advertisement display data from the advertisement DB server 14, it displays an advertisement based on the received advertisement display data on the display device 11a (step ST38).

[0071] When the control unit 41 of the automatic ticket gate 11 executes the display of the advertisement, it transmits data indicating the displayed advertisement as advertisement display history information to the recommendation information server 13 in association with the user's ID (step ST39).

[0072] The recommendation information server 13 receives information indicating the advertisement display history from the automatic ticket gate 11 via the communication interface 35. When the processor 31 of the recommendation information server 13 receives the advertisement display history from the automatic ticket gate 11, it adds (updates) the information indicating the received advertisement to the advertisement history information included in the user data of the user (step ST40).

[0073] Next, a process of updating the user database 34b by the recommended information server 13 in the information recommendation system 1 according to the embodiment will be described. FIG. 6 is a flowchart for explaining an example of an update process of the user database 34b by the recommended information server 13 in the information recommendation system 1 according to the embodiment. In the information recommendation system 1, the recommended information server 13 receives log data such as ticket gate processing results from the automatic ticket gate 11 (step ST51). The log data includes information that associates the ticket gate processing results, including the entry or exit location (station name, ticket gate name) and the date and time of passage, with the user's identification information (ID).

[0074] When the processor 31 of the recommendation information server 13 receives the log data from the automatic ticket gate 11, the processor 31 stores the received log data in the storage unit 34 (step ST51). Here, the processor 31 stores movement history information for each user in the storage unit 34 based on information indicating the ticket gate processing result included in the log data.

[0075] FIG. 7 is a diagram showing an example of movement history information based on log data acquired from the automatic ticket gate 11. As shown in FIG. In the example shown in Figure 7, movement history information is generated for each user and includes the name of the entry station, the name of the exit station, the time of exit, and information about the exit station (neighborhood information) as movement history that can be identified from the log data.

[0076] The entrance station name, exit station name, and exit time are identified from the ticket gate processing result obtained from the automatic ticket gate 11. The entrance station name is identified from log data including the entrance information of the user. The exit station name and exit time are identified from the ticket gate processing result (log data) including the exit information of the user, and are stored in association with the entrance station name of the corresponding entrance information.

[0077] The neighborhood information is information indicating content expected from a station. The neighborhood information is identified based on station surrounding information that is set for each station and indicates facilities and events around the station. The station surrounding information may be stored in the storage unit 34 of the recommended information server 13, or may be information collected from an external server, etc.

[0078] After saving the movement history information as log data from the automatic ticket gate 11, the processor 31 of the recommendation information server 13 extracts characteristic movements that indicate the user's preferences from the movement history information of the user (step ST53). For example, the processor 31 extracts multiple movements to a station where an aquarium is located, movements to the nearest station to a game venue on the day of a specific sports match, movements to the nearest station to a live music venue when a live music concert is held, etc.

[0079] When the processor 31 of the recommended information server 13 extracts the characteristic travel history, it estimates information (preferred content) that is likely to interest the user based on the extracted travel history (step ST54). For example, the processor 31 lists content that is likely to interest the user by matching the travel locations and travel dates and times identified from the extracted travel history with information such as events.

[0080] When the processor 31 of the recommended information server 13 estimates the user's favorite content, it updates the favorite information pool in the user data of the user using the favorite content estimated from the log data (step ST55). The favorite information pool is information that lists the content that the user likes (favorite content), and is stored in the user database 34b as part of the user data in association with the user's ID.

[0081] FIG. 8 is a diagram showing an example of a preference information pool for each user.

[0082] As shown in Fig. 8, a preference information pool is provided for each user and stores information indicating each user's preference content. The preference information pool may be any information that lists the preference content of each user, and is information that indicates a plurality of preference contents for each user, as shown in Fig. 8. For example, when processor 31 estimates a user's preference content, it updates the preference information pool for each user as shown in Fig. 8 by adding or overwriting the preference content estimated from the log data.

[0083] When the processor 31 of the recommendation information server 13 updates the user's preference information pool, it updates the user's preference vector according to the content in the preference information pool (step ST56). The preference vector indicates multiple types of feature quantities. For example, the preference vector is a feature vector that indicates multiple types of feature quantities obtained by breaking down each user's preference content into numerical values ​​for multiple feature quantities and integrating these values. The preference vector is calculated as a feature quantity vector obtained by general methods such as machine learning and factorization.

[0084] FIG. 9 is a diagram showing an example of a preference vector for each user. As shown in Figure 9, a preference vector is provided for each user and indicates the user's preferences using multiple types of feature quantities. The preference vector is calculated from the user's preference content group. Therefore, each user's preference vector is updated when the user's preference information pool is updated.

[0085] Through the above-described processing, the recommendation information server updates the preference information pool and preference vector contained in the user data of each user stored in the user database. The preference information pool may be updated based on preference content estimated from purchase data, search data, or viewing data. For example, if information (purchase data) such as payment history using an IC card, a credit card, or a smartphone or other information terminal can be obtained, the recommendation information server may analyze the purchasing trends of each user from the purchasing data and estimate the content (preferred content) that each user is interested in from the purchasing trends of each user.

[0086] In addition, if information indicating the search history on the Web when a user is identified can be obtained, the recommended information server may estimate the preferred content for each user by identifying information related to the search history on the Web. In addition, if viewing data indicating the viewing history of video sites or subscription services when a user is identified can be obtained, the recommendation information server may estimate the preferred content of each user based on the viewing data of each user.

[0087] Next, a process (advertisement instruction process) in which the recommendation determination server 12 in the information recommendation system 1 according to the embodiment instructs the automatic ticket gate 11 which advertisement to display will be described. FIG. 10 is a flowchart for explaining an example of processing (advertisement instruction processing) in which the recommended information server 13 in the information recommendation system 1 according to the embodiment instructs an advertisement to be displayed on the automatic ticket gate 11. In the information recommendation system 1, the recommendation determination server 12 acquires the ID of the user who will perform the ticket gate process from the automatic ticket gate 11 (step ST61). When the processor 21 of the recommendation determination server 12 receives the ID from the automatic ticket gate 11, it requests the recommendation information server 13 for the user data and advertisement display history of the user identified by the received ID (step ST62).

[0088] When the processor 21 of the recommendation determination server 12 acquires the user data and advertisement display history of the user from the recommendation information server 13, the processor 21 performs a reference user selection process to select users to be referenced in order to determine the advertisement to be provided to the user, based on the user data of the user (step ST63). As the reference user selection process, the processor 21 creates a reference user list that lists reference users using the feature amounts of the user data of the user. An example of the reference user selection process will be described in detail later.

[0089] After generating the reference user list, the processor 21 of the recommendation determination server 12 executes a ranking process to rank each content in order to select candidate advertisements to be provided to the user using the content preferences of the reference user shown in the reference user list (step ST64). In the ranking process, the processor 21 creates a ranking of the content preferences of each user in the reference user list in order of the degree of deviation from the user's own preferences. An example of the ranking process will be described in detail later.

[0090] After ranking the content preferences of the referring user, the processor 21 of the recommendation determination server 12 acquires advertising data (data indicating the details of the advertisements) related to each content in descending order of rank from the advertisement DB server 14 (step ST65). For example, the processor 21 acquires a predetermined number of advertising data for each content in descending order of rank from the advertisement DB server 14.

[0091] When processor 21 acquires multiple pieces of advertisement data related to highly ranked content, it performs advertisement determination processing to determine one advertisement (display advertisement) to be displayed on automatic ticket gate 11 from the advertisement data (step ST66). For example, processor 21 of recommendation determination server 12 determines, as the advertisement to be displayed, the advertisement data that is highest in the order according to the rank of the content among the advertisement data with the smallest display history among the acquired advertisement data. An example of the advertisement determination processing will be described in detail later.

[0092] When the processor 21 of the recommendation determination server 12 determines the display advertisement, it transmits a display instruction for the display advertisement to the automatic ticket gate 11 (step ST67). As a result, the automatic ticket gate 11 obtains the display data of the display advertisement specified by the recommendation determination server 12 from the advertisement DB server 14 and displays it on the display device 11a.

[0093] Next, a process of selecting a referring user by the recommendation determination server 12 in the information recommendation system 1 according to the embodiment will be described. FIG. 11 is a flowchart illustrating an example of a process for selecting a referring user by the recommendation determination server 12 in the information recommendation system 1 according to the embodiment. The processor 21 of the recommendation determination server 12 selects a user to refer to by focusing on the order of multiple types of feature amounts for the user's preference vector acquired from the recommendation information server 13. The processor 21 first sets a variable i for sequentially specifying multiple types of feature amounts to an initial value (step ST71). The processor 21 selects the i-th feature amount from the user data of the user (step ST72).

[0094] When the processor 21 selects the ith feature, it adds other users similar to the ith feature of the user (subject) to the reference user list (step ST73). For example, the processor 21 specifies the ith feature of the subject as a search key, and requests a predetermined number of other users similar to the ith feature of the subject from the recommended information server 13. The processor 21 acquires information indicating the predetermined number of other users similar to the ith feature of the subject searched by the recommended information server 13, and adds the information to the reference user list.

[0095] When the processor 21 adds a predetermined number of other users who are similar to the i-th feature of the subject to the reference user list, the processor 21 calculates the similarity of the feature of each other user to the subject in the reference user list to which the other users have been added (step ST74). For example, the processor 21 calculates COS similarity as the similarity of the other users to the subject for feature quantities other than the i-th feature quantity. The COS similarity indicates the similarity for multiple types of feature quantities and is expressed as a value from "1" to "-1." The COS similarity indicates that "1" is the most similar (close), and "-1" indicates that the least similar (distant). Therefore, the processor 21 can determine that the closer the COS similarity is to "-1," the lower the similarity to the subject.

[0096] After calculating the similarity of other users to the feature amount of the target person, processor 21 updates the reference user list by leaving a predetermined number of other users in the reference user list in order of decreasing similarity (step ST75). In other words, processor 21 deletes a predetermined number of other users from the reference user list in order of decreasing similarity to the feature amount of the target person.

[0097] Fig. 12 is a diagram showing an example in which four other users T, O, S, and H, whose feature 1 (feature amount i=1) is similar to that of subject A, are added to the reference user list. Fig. 13 is a diagram showing an example in which two other users O and H are left from the reference user list shown in Fig. 12 in order of COS similarity closest to "-1."

[0098] In the example shown in FIG. 12, other users T, O, S, and H whose difference (error) with respect to the feature amount of feature 1 (first feature amount) of subject A is within 0.1 are added to the reference user list. The processor 21 calculates the COS similarity of the other users to subject A for multiple feature amounts other than the first feature amount (feature amount 1). The example shown in FIG. 12 shows an example in which the COS similarity of other users T, O, S, and H to subject A is calculated for multiple feature amounts other than feature amount 1 (feature amounts of features 2-4).

[0099] When the processor 21 calculates the COS similarity shown in Fig. 12, it leaves two other users O and H whose COS similarity is close to -1 in the reference user list, and deletes the remaining users T and S from the reference user list, as shown in Fig. 13. As a result, the recommendation determination server 12 creates the list shown in Fig. 13 as a reference user list created (updated) focusing on feature 1 of the target person A.

[0100] When processor 21 updates the reference user list focusing on the i-th feature, it increments variable i (i=i+1) (step ST76). If processor 21 has not completed updating the reference user list focusing on all feature amounts (step ST77, NO), it returns to step ST72 and updates the reference user list focusing on the updated i(i+1)-th feature.

[0101] Fig. 14 is a diagram showing an example in which two other users I and B who are similar to subject A in terms of feature 2 (feature amount i=2) are added to the reference user list shown in Fig. 13. Fig. 15 is a diagram showing an example in which two other users O and H are left from the reference user list shown in Fig. 14 in order of COS similarity closest to "-1".

[0102] In the example shown in Fig. 14, other users I and B whose difference (error) with respect to the feature amount of feature 2 (second feature amount) of subject A is within 0.1 are added to the reference user list in Fig. 13. Processor 21 calculates the COS similarity of other users to the subject A for multiple feature amounts other than the second feature amount (feature amount 2). Fig. 14 shows an example in which the COS similarity of other users I and B to subject A is calculated for multiple feature amounts other than feature amount 2 (feature amounts of features 1, 3-4).

[0103] When the processor 21 calculates the COS similarity shown in Fig. 14, it leaves two other users H and I whose COS similarity is close to -1 in the reference user list, and deletes the remaining users O and B from the reference user list, as shown in Fig. 15. As a result, the recommendation determination server 12 creates the list shown in Fig. 15 as a reference user list created (updated) focusing on feature 2 of the target person A.

[0104] According to the above processing, the recommendation determination server 12 can create a reference user list by repeatedly updating the reference user list, focusing on all features in turn and leaving other users with low similarity to the target user.

[0105] Next, a process of selecting a referring user by the recommendation determination server 12 in the information recommendation system 1 according to the embodiment will be described. FIG. 16 is a flowchart illustrating an example of a content ranking process performed by the recommendation determination server 12 in the information recommendation system 1 according to the embodiment. As the ranking process, the processor 21 of the recommendation determination server 12 executes a process of ranking the content extracted from the content preferences of each user in the referring user list.

[0106] The processor 21 first refers to the preference information pool (preference content group) of each user listed in the reference user list (step ST81). For example, the processor 21 acquires preference content groups in the preference information pools of other users in the reference user list from the recommended information server 13.

[0107] When processor 21 acquires the group of preference contents of other users, processor 21 extracts content (recommended content) that is not included in the preference information pool of the target user from the group of preference contents of other users (step ST82). Here, the recommended content indicates candidate content for advertisement to be displayed.

[0108] When processor 21 extracts content not included in the target user's preference information pool from the group of other users' preference contents, processor 21 calculates the degree of deviation of each extracted content from the target user's preference (step ST83). For example, processor 21 calculates the degree of deviation between the feature of the target user's preference, which is made up of multiple types of feature, and the feature of the content, which is made up of multiple types of feature, for each piece of content.

[0109] After calculating the degree of deviation between the subject's preferences and each piece of content, processor 21 ranks each piece of content in order of deviation from the subject's preferences based on the calculated degree of deviation for each piece of content (step ST84). For example, processor 21 ranks each piece of content by ranking a plurality of pieces of content extracted from other users' preference information pools in order of deviation from the subject's preferences.

[0110] FIG. 17 is a diagram showing an example of a preference information pool (preference content group) of subject A and an example of a preference information pool (preference content group) of other users H and I. FIG. 18 is a diagram showing an example of a content group (content list) extracted from the preference information pool of subject A shown in FIG. 17 and the preference information pool of other users H and I. FIG. 19 is a diagram showing an example of the feature amount of subject A and the feature amount of each content shown in FIG. 18. FIG. 20 is a diagram showing an example of ranking each content (ranking the content) based on the feature amount of subject A shown in FIG. 19 and the feature amount of each content.

[0111] The processor 21 acquires a group of preference contents from the preference information pool of other users listed in the reference user list from the recommended information server 13. As a result, the processor 21 acquires a group of preference contents of the target user and a group of preference contents of other users listed in the reference user list as shown in FIG.

[0112] 17, processor 21 selects up to three pieces of content from each other user's preference content group that are not included in the target user's preference information pool as recommended content. If the other user's preference content group contains three or more pieces of content that are not included in the target user's preference information pool, three pieces of content may be randomly selected from those pieces of preference content.

[0113] The example shown in Fig. 18 shows an example in which recommended content is selected from the favorite content of other users for subject A shown in Fig. 17. Specifically, the recommended content for subject A shown in Fig. 18 is an example in which "aquarium" and "cafe" are selected from the favorite content group of other user H shown in Fig. 17, and "railway," "painting," and "comedy" are selected from the favorite content of user I shown in Fig. 17.

[0114] After selecting the recommended contents as shown in Fig. 18, processor 21 calculates the COS similarity indicating the similarity in taste to subject A for each recommended content, as shown in Fig. 19. The COS similarity of each recommended content is calculated from multiple types of feature amounts forming the subject's preference vector and multiple types of feature amounts of the recommended content.

[0115] After calculating the COS similarity of each recommended content as shown in Fig. 19, the processor 21 ranks them in order as shown in Fig. 20. The ranking of recommended content shown in Fig. 20 is created based on the COS similarity of each recommended content shown in Fig. 19. Specifically, in the example shown in Fig. 20, "Aquarium", which has the COS similarity closest to "-1", is ranked first, "Railway", which has the COS similarity next to "-1" after "Aquarium", is ranked second, and "Cafe", which has the COS similarity next to "-1" after "Railway", is ranked third.

[0116] Next, an advertisement determination process performed by the recommendation determination server 12 in the information recommendation system 1 according to the embodiment will be described. FIG. 21 is a flowchart illustrating an example of an advertisement determination process by the recommendation determination server 12 in the information recommendation system 1 according to the embodiment. As an advertisement determination process, the processor 21 of the recommendation determination server 12 determines one advertisement to be displayed at the automatic ticket gate 11 from advertisements related to the ranked recommended content.

[0117] That is, processor 21 refers to the ranking of the recommended content (step ST91), and searches for advertisements (advertisements to be displayed candidates) related to keywords indicating the content in order from the top-ranked recommended content (step ST92). For example, processor 21 acquires advertisement data indicating advertisements (advertisements to be displayed candidates) searched for using keywords indicating each recommended content in order from the top-ranked one as search keys from advertisement DB server 14. Here, advertisements to be displayed candidates may be ranked according to the ranking of the recommended content.

[0118] When processor 21 acquires the advertisement data indicating a group of advertisements related to the top-ranked recommended content, processor 21 acquires the number of times each advertisement indicated by the advertisement data has been displayed to the target person (step ST93). For example, processor 21 searches the advertisement display history of each advertisement indicated by the advertisement data from the advertisement display history of the target person, and identifies the number of times each advertisement has been displayed to the target person based on the searched display history.

[0119] After acquiring the advertisements to be displayed and the display history of each advertisement, processor 21 determines one advertisement to be displayed on automatic ticket gate 11 (step ST94). Processor 21 determines the advertisement with the smallest number of times of display among the advertisements to be displayed as candidates searched for based on the recommended content as the advertisement to be displayed on automatic ticket gate 11. Furthermore, if there are multiple advertisements with the smallest number of times of display, processor 21 may determine one advertisement based on the ranking of the recommended content that was the search source for each advertisement.

[0120] Fig. 22 is a diagram showing examples of keywords set for each advertisement managed by the advertisement DB server 14. Fig. 23 is a diagram showing examples of advertisements that are display candidates selected from the recommended content shown in Fig. 20 and the keywords of each advertisement shown in Fig. 22. Fig. 24 is a diagram showing a display example of an advertisement displayed on the display 11a of the automatic ticket gate 11.

[0121] As shown in Fig. 22, when registering an advertisement, the advertisement DB server 14 registers keywords related to the advertisement to be registered along with data for displaying the advertisement (advertisement display data). The keywords are words corresponding to content selected by the recommendation determination server 12 and are used as search keys for searching for advertisements. The advertisement DB server 14 manages not only the advertisement display data but also information indicating the content of the advertisement (for example, the advertisement title, etc.) and keywords related to the advertisement as advertisement data, as shown in Fig. 22.

[0122] The information recommendation system according to this embodiment is intended to display advertisements on the display unit of an automatic ticket gate. Therefore, the advertisement DB server is expected to register advertisements related to "information on businesses operated by railway operators," "tourist information around stations," "event information around stations," "information on revitalization projects around stations," "experience information at facilities around stations," and so on.

[0123] As specific examples, "information on businesses operated by railway operators" could include information on boxed lunches and side dishes on sale within stations. "Tourist information around stations" could include information on tourist attractions around stations, such as zoos. "Event information around stations" could include sales information at shopping arcades around stations. "Experience information at facilities around stations" could include information on tickets to gymnasiums around stations.

[0124] 20 is obtained, the processor 21 requests advertisement data of advertisements searched for using keywords corresponding to each recommended content as search keys from the advertisement DB server 14. For example, the processor 21 requests two advertisements searched for using keywords for each recommended content.

[0125] In the example shown in Fig. 23, two advertisements related to "Aquarium", which is the recommended content ranked first, two advertisements related to "Railway", which is the recommended content ranked second, and two advertisements related to "Cafe", which is the recommended content ranked third, are obtained. In the advertisement list shown in Fig. 23, the two advertisements related to "Aquarium", the two advertisements related to "Railway", and the two advertisements related to "Cafe" are listed in order.

[0126] Furthermore, when the advertisement list is obtained using the keywords of each recommended content, the processor 21 identifies the number of times each advertisement in the advertisement list has been displayed to the target user A. The processor 21 identifies the number of times each advertisement has been displayed to the target user A from the advertisement display history included in the user data of the target user A. After identifying the number of times each advertisement in the advertisement list has been displayed to the target user A, the processor 21 associates the number of times each advertisement has been displayed with each advertisement in the advertisement list, as shown in FIG. 23 .

[0127] After generating an advertisement list indicating the number of times of display, the processor 21 identifies one advertisement with the smallest number of times of display from the top of the list. For example, when the advertisement list shown in Fig. 23 is obtained, the processor 21 determines the advertisement "□□ Railway Special Offer Tickets on Sale Now" which is at the top and has a number of times of display of 0 as the advertisement to be displayed on the automatic ticket gate 11. This allows the processor 21 of the recommendation determination server 12 to display an advertisement such as that shown in Fig. 24 on the display 11a of the automatic ticket gate 11.

[0128] Next, a process of updating the user database 34b by the recommended information server 13 in the information recommendation system 1 according to the embodiment will be described. FIG. 25 is a flowchart for explaining an example of an update process of the user database 34b by the recommended information server 13 in the information recommendation system 1 according to the embodiment. Here, in the information recommendation system 1, the recommendation information server 13 is assumed to be communicatively connected to an external payment system via a communication interface 35. In the payment process, the external payment system acquires the identification information of the person making the payment (user) along with information indicating the payment details. Here, the external payment system is assumed to be able to acquire, as the user's identification information, a user ID registered in the user database 34b or identification information linked to the user ID. For example, the external payment system performs payment processing for goods and services using a boarding tool such as an IC card, a credit card, or a QR code (registered trademark).

[0129] The processor 31 of the recommendation information server 13 acquires information related to payment (payment information) from an external payment system connected via the communication interface 35 (step ST101). The payment information acquired by the recommendation information server 13 may be information that includes information that can identify the person making the payment and the details of the product or service for which payment has been made (payment details). Here, the payment information includes a user ID as identification information of the person making the payment, and information indicating the product or service for which payment has been made as information indicating the payment details.

[0130] The processor 31 of the recommendation information server 13 identifies the user who made the payment by comparing the user ID included in the payment information with the users registered in the user database 34b. If the processor 31 of the recommendation information server 13 can identify the user who made the payment, it identifies the advertisement display history indicating advertisements that have been displayed to the user in the past (step ST102). Note that if the processor 31 of the recommendation information server 13 cannot identify the user from the payment information, it may end the process of updating the user database.

[0131] The processor 31 of the recommendation information server 13 determines whether the payment details are related to a previously displayed advertisement (step ST103). For example, the processor 31 identifies a product or service for which payment has been made from the payment information, and determines whether an advertisement related to the product or service for which payment has been made has been displayed to the user.

[0132] When the processor 31 of the recommended information server 13 determines that the payment details are related to a previously displayed advertisement (step ST103, YES), it determines (step ST104) the effectiveness of the information (advertisement) recommended by the advertisement display process (information recommendation process) of the information recommendation system 1. For example, the processor 31 determines whether the information (advertisement) recommended by the information recommendation process was effective (whether the advertisement led to the user's consumption behavior) depending on whether the displayed advertisement related to the payment details is an advertisement selected by the advertisement display process.

[0133] Furthermore, when the processor 31 of the recommended information server 13 can identify the payment details and the user from the payment information, it updates the preference information pool of the user according to the preferences of the user estimated from the payment details (step ST105). For example, the processor 31 of the recommended information server 13 estimates information (preference content) that is estimated to be of interest to the user based on the payment details included in the payment information. The processor 31 updates the preference information pool of the user with the preference content estimated from the payment details, thereby updating the user database 34b.

[0134] Next, a process of updating the user database 34b by the recommended information server 13 in the information recommendation system 1 according to the embodiment will be described. FIG. 26 is a flowchart for explaining an example of an update process of the user database 34b by the recommended information server 13 in the information recommendation system 1 according to the embodiment. Here, in the information recommendation system 1, the recommended information server 13 is assumed to be communicatively connected to an external facility management system and a point management system via a communication interface 35.

[0135] The external facility management system acquires facility usage information including information indicating the content of facility usage and user (user) identification information as part of a usage management process for managing facility usage. Here, the external facility management system acquires, as user identification information, a user ID or identification information linked to a user ID registered in the user database 34b. For example, the external facility management system identifies a user of a facility or a specific service using a boarding tool such as an IC card, credit card, or QR code, and performs usage management process for the facility or specific service by that user.

[0136] The point management system is a system that manages points that can be used to purchase various services and goods, including the use of facilities. The points managed by the point management system are numerical values ​​that can be used as part or all of the payment for using various services and purchasing goods. The point management system manages the number of points for each user and uses (consumes) the points according to the user's intentions. The point management system also works in conjunction with various systems, including the information recommendation system 1, and accumulates points for specific users in response to requests for point allocation from other systems.

[0137] The processor 31 of the recommendation information server 13 acquires information related to facility use (facility use information) from an external facility management system connected via the communication interface 35 (step ST111). The facility use information acquired by the recommendation information server 13 may be information that can identify the facility or service used, and includes the details of the use (use details) and user identification information. Here, the facility use information includes a user ID as identification information for the facility user, and includes information indicating the facility or service used as information indicating the use details.

[0138] The processor 31 of the recommendation information server 13 identifies the user who used the facility by comparing the user ID included in the facility usage information with the user ID registered in the user database 34b. When the processor 31 of the recommendation information server 13 identifies the user who used the facility, it identifies the advertisement display history indicating advertisements that have been displayed to the user in the past (step ST112). Note that if the processor 31 of the recommendation information server 13 cannot identify the user from the facility usage information, it may end the process of updating the user database.

[0139] The processor 31 of the recommendation information server 13 determines whether the usage details included in the facility usage information are related to an advertisement that has been displayed to the user in the past (step ST113). For example, the processor 31 identifies a facility or a used service from the facility usage information, and determines whether an advertisement related to the used facility or service has been displayed to the user in the past.

[0140] When the processor 31 of the recommended information server 13 determines that the usage content is related to a previously displayed advertisement (step ST113, YES), it determines (step ST114) the effectiveness of the information (advertisement) recommendation by the advertisement display process (information recommendation process) of the information recommendation system 1. For example, the processor 31 determines whether the information (advertisement) recommended by the above-mentioned information recommendation process was effective (whether the advertisement led to the user's behavior) depending on whether the displayed advertisement related to the usage content is an advertisement selected by the above-mentioned advertisement display process.

[0141] When the processor 31 of the recommended information server 13 determines the effectiveness of the information recommendation, it performs a process of awarding points to the user who is a facility user, according to the usage content based on the facility usage information (step ST115). For example, the processor 31 calculates the number of points to be awarded to the user from the usage content and requests an external point management system to award the calculated number of points to the user. In response, the external point management system awards points to the facility user according to the advertisement provided (recommended) by the information recommendation system 1. Furthermore, the number of points calculated by the processor 31 (the number of points to be awarded to the user) may be set for each facility or service, or may be based on the effectiveness of the information recommendation. By awarding such points, it is possible to encourage users to use facilities or services prompted by the information (advertisement) recommended by the information recommendation system 1.

[0142] Furthermore, when the processor 31 of the recommended information server 13 can identify the facility use details and the user from the facility use information, it updates the preference information pool of the user according to the preferences of the user estimated from the use details (step ST116). For example, the processor 31 of the recommended information server 13 estimates information (preference content) that is estimated to be of interest to the user based on the use details included in the facility use information. The processor 31 updates the preference information pool of the user with the preference content estimated from the use details, thereby updating the user database 34b.

[0143] As described above, the information recommendation system according to the embodiment acquires information about a user's preferences, extracts recommended content from the preference content of a reference user who has different preferences from the user, and determines an advertisement to be displayed on a display device from the display candidate advertisements searched based on the recommended content.

[0144] This allows the system to select (recommend) advertisements to be displayed to users by referencing the preference information of other users who have similar but different preferences to the user. As a result, advertisements that may lead to new discoveries that would not have been possible based on the user's own preferences alone can be displayed to the user.

[0145] Furthermore, according to the information recommendation system of the embodiment, information in fields that do not fit the user's preferences can be presented to the user as advertisements. Therefore, even for users who are not interested in ordinary advertisements based on their own preferences, novel information that can lead to new discoveries can be presented, thereby encouraging new behavior from the user.

[0146] Furthermore, according to the information recommendation system of the embodiment, it is possible to select (recommend) advertisements to be displayed to a user using an automatic ticket gate from the preference content of other users who have different preferences from the preferences estimated from the user's travel history, etc. As a result, it is possible to suggest new experiences associated with travel to users of railways, etc., and encourage a wide range of actions that are different from usual, thereby stimulating user activity.

[0147] Furthermore, according to the information recommendation system of the embodiment, it is possible to present information (advertisements) at the automatic ticket gates that encourages users to use facilities around the station from a perspective different from the user's own preferences, which is expected to lead to an increase in the number of users of facilities around the station. As a result, it is expected that the value of the area around the station will be improved as the number of users of facilities around the station increases.

[0148] Furthermore, the information recommendation system according to the embodiment can encourage consumption behavior based on preferences that differ from those of individual users, and therefore, by incorporating information about the lifestyle service businesses of railway operators into the information provided as advertisements, it is expected that this will lead to an increase in direct revenue for the railway operators.

[0149] Furthermore, according to the information recommendation system of the embodiment, advertisers who request advertisements can widely promote their advertisements to new customer segments that cannot be reached by existing advertising systems, and it is expected that the publicity effect of even existing advertisements will be increased.

[0150] Although several embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be embodied in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their modifications are included within the scope and spirit of the invention, and are also included in the scope of the invention and its equivalents as defined in the claims. [Explanation of symbols]

[0151] 1...information recommendation system, 11...automatic ticket gate, 11a...display (display device), 12...recommendation determination server (recommendation determination device), 13...recommendation information server, 14...advertising DB server, 21...processor, 21a...user search unit, 21b...ranking unit, 21c...advertising determination unit, 24...memory unit, 25...communication interface, 31...processor, 31a...feature calculation unit, 34...memory unit, 34b...user database, 35...communication interface, 41...control unit, 42...data memory, 43...communication unit, 44...ID reader, 45...passage control unit.

Claims

1. A recommendation determination device that determines display information to be displayed on a display device, a communication unit that communicates with the display device; a processor that acquires information about preferences of a target person identified by identification information acquired from a person visible to the display device, and supplies to the display device a display instruction for one piece of display information determined from display information that is a display candidate searched based on recommended content extracted from preference content of a reference user having a preference different from that of the target person; A recommendation determination device having the above.

2. the communication unit communicates with an automatic ticket gate equipped with the display device, the processor acquires identification information of a user who performs a ticket examination process at the automated ticket gate, and provides the automated ticket gate with an instruction to display one advertisement determined from advertisement display candidates searched based on the recommended content by the display device; The recommendation determination device according to claim 1 .

3. The information about the subject's preferences is estimated based on a movement history identified from a ticket gate processing result by an automatic ticket gate. The recommendation determination device according to claim 2 .

4. The processor acquires a plurality of types of feature quantities that indicate the preferences of the user, and selects, as a reference user, another user who has a plurality of types of feature quantities that are similar to the user in one of the feature quantities and deviate from the user in other feature quantities. The recommendation determination device according to claim 2 .

5. the processor determines a degree of discrepancy based on a COS similarity calculated from the feature amount of the user and the feature amount of the other user, excluding the one similar feature amount. The recommendation determination device according to claim 4 .

6. The processor extracts, from the content preferences of the referring user, content that is not included in the preferences of the user as recommended content, ranks each of the extracted recommended content in order of the degree of deviation from the preferences of the user, and determines one advertisement from among the display candidates of advertisements searched based on the recommended content with the highest rank. The recommendation determination device according to claim 2 .

7. the processor determines a degree of deviation of the recommended content from the user's preference based on a COS similarity calculated from a feature amount indicating the user's preference and a feature amount of the recommended content; The recommendation determination device according to claim 6 .

8. the processor acquires a display history indicating the number of times an advertisement has been displayed to the user, and determines one advertisement that has been displayed the least number of times from among display candidates for advertisements searched for based on the recommended content with the highest ranking; The recommendation determination device according to claim 6 .

9. A display information recommendation method for determining display information to be displayed on a display device, comprising: identifying a target person based on identification information obtained from a person visible to the display device; obtaining information regarding the subject's preferences; extracting recommended content from the preference content of a reference user who has a preference different from that of the target user; determining one piece of display information from the display information that is a display candidate searched based on the recommended content; supplying a display instruction for the determined piece of display information to the display device; How to recommend information to display.

10. On the computer, identifying the target person using identification information obtained from the person visible to the display device; obtaining information regarding the subject's preferences; extracting recommended content from the preference content of a reference user who has a preference different from that of the target user; determining one piece of display information from the display information that is a display candidate searched based on the recommended content; supplying a display instruction for the determined piece of display information to the display device; A program that makes it happen.

Images