JP7453120B2 - Selection device, selection method and selection program - Google Patents

Description

The present invention relates to a selection device, a selection method, and a selection program.

Conventionally, techniques for providing various information to users are known. For example, there is a technology that uses a recommendation system to select content to be provided to a user based on the content viewed by the user.

Japanese Patent Application Publication No. 2016-051357

However, in the above-mentioned conventional technology, the content to be provided to the user is only selected based on the user's content viewing using a recommendation system, and it is not possible to appropriately select information other than the content. is not limited.

The present application has been made in view of the above, and aims to provide a selection device, a selection method, and a selection program that can appropriately select transaction target information to be provided to a user.

The selection device according to the present application includes a first selection unit that selects a transaction object candidate that is a candidate for a transaction object to be provided to the user from among the transaction objects based on user information regarding the user; When a person displays a predetermined content, the transaction target displayed in the predetermined content is selected from among the transaction target candidates selected by the first selection unit according to the display number of times the predetermined content has been displayed. and a second selection unit that makes a selection based on the information.

According to one aspect of the embodiment, it is possible to appropriately select transaction target information to be provided to a user.

FIG. 1 is a diagram illustrating an example of a selection process executed by a selection device according to an embodiment. FIG. 2 is a diagram illustrating a specific example of the ranking of transaction targets according to the embodiment. FIG. 3 is a diagram showing a specific example of content according to the embodiment. FIG. 4 is a diagram illustrating a specific example of the ranking of transaction target candidates according to the embodiment. FIG. 5 is a diagram illustrating a configuration example of a selection system according to an embodiment. FIG. 6 is a diagram illustrating an example of a user information storage unit according to the embodiment. FIG. 7 is a diagram illustrating an example of a transaction target information storage unit according to the embodiment. FIG. 8 is a diagram illustrating an example of a content storage unit according to the embodiment. FIG. 9 is a flowchart illustrating an example of the flow of selection processing executed by the selection device according to the embodiment. FIG. 10 is a hardware configuration diagram showing an example of a computer that implements the functions of the selection device.

DESCRIPTION OF THE PREFERRED EMBODIMENTS Below, a selection device, a selection method, and a mode for implementing a selection program (hereinafter referred to as an "embodiment") according to the present application will be described in detail with reference to the drawings. Note that the selection device, selection method, and selection program according to the present application are not limited to this embodiment. Moreover, each embodiment can be combined as appropriate within the range that does not conflict with the processing contents. Further, in each of the embodiments below, the same parts are given the same reference numerals, and redundant explanations will be omitted.

[1. Example of provision processing indicated by selection device]
First, an example of the selection process executed by the selection device 100 will be described using FIG. 1. FIG. 1 is a diagram illustrating an example of a selection process executed by a selection device 100 according to an embodiment. FIG. 1 shows a selection process for a transaction object to be traded at an auction site or a flea market site. Here, the transaction object is the use of a product or service.

First, an example will be described in which content including a transaction object with a high score based on user information is provided. As shown in FIG. 1, the selection device 100 acquires user information regarding the user from the external device 20 (step S1).

In the example of FIG. 1, it is assumed that the user has registered as a member of a predetermined service. Furthermore, when registering as a member, the user shall register attribute information regarding attributes such as the user's gender, age, telephone number, and address. In this case, the selection device 100 acquires, as user information, user attribute information, purchase history on Internet shopping, online shopping malls, auction sites, flea market sites, etc., content display history, etc. from the external device 20. do. For example, the selection device 100 acquires user information from the external device 20 every predetermined period (eg, 5 hours, 1 day, etc.).

Further, the selection device 100 acquires information regarding the transaction object from the external device 20 (step S2). For example, the selection device 100 acquires information regarding a transaction object traded at an auction site or a flea market site from the external device 20. To give a more specific example, the selection device 100 acquires information regarding a smartphone being bid on at an auction site from the external device 20. For example, the selection device 100 acquires information regarding the transaction object from the external device 20 at predetermined intervals.

In the above example, the selection device 100 performs the acquisition process separately in step S1 and step S2, but the selection device 100 receives user information from the external device 20 and information regarding the transaction The information may also be acquired at the same time.

Subsequently, the selection device 100 receives a request for content provision from the terminal device 10 (step S3). For example, the selection device 100 receives a request to provide a portal site from the terminal device 10 as an example of content.

Note that in the above example, the selection device 100 performs step S3 after performing step S1 and step S2, but the selection device 100 performs step S1 after performing step S3. and step S2 may be executed.

Then, the selection device 100 calculates the first score SC1 of the selected transaction object (step S4). For example, the selection device 100 uses conventional technology such as machine learning to create a learning model that outputs a first score SC1 indicating the possibility of purchasing the transaction object by inputting the transaction object and user information. generate.

For example, the selection device 100 generates a learning model that has learned the relationship between a transaction object, user information, and information regarding whether the transaction object was purchased by the user in the past. Thereby, the selection device 100 calculates the first score SC1 by inputting the transaction object and the user information into the learning model.

Further, the selection device 100 acquires, as learning data, information indicating various types of user information (attributes, etc.) and information on a transaction target selected by the user. When the selection device 100 inputs the user information and the information of the transaction object selected by the user, the selection device 100 inputs the user information and the information of the transaction object not selected by the user. The learning model is trained to output a higher score than when In addition, various other techniques can be employed to quantify the extent to which a user shows interest in a transaction object.

For example, the selection device 100 uses deep learning to perform a learning model generation process using a method such as backpropagation (error backpropagation method). For example, the selection device 100 uses a learning process to adjust the weight values that are taken into consideration when transmitting values between nodes. In this way, the selection device 100 generates a learning model through a process such as backpropagation that corrects the weight value so that the error between the output of the learning model and the correct answer corresponding to the input is reduced. For example, the selection device 100 generates a learning model by performing processing such as backpropagation to minimize a predetermined loss function. Thereby, the selection device 100 can perform the learning model generation process. Note that the calculation method is not limited to the above example, and various calculation methods can be adopted based on other machine learning.

Here, using FIG. 2, a specific example of the ranking based on the first score SC1 of the transaction object will be shown. FIG. 2 is a diagram illustrating a specific example of the ranking of transaction targets according to the embodiment. In the example of FIG. 2, a table TA1 is shown in which the ranking of the transaction object, the transaction object ID (Identifier) which is identification information for identifying the transaction object, and the first score SC1 are associated with each other. Here, it is assumed that table TA1 shows first scores SC1 for 20 transaction objects among the 50 transaction objects acquired from external device 20. Note that in the example of FIG. 2, table TA1 shows first scores SC1 for ten transaction targets.

According to table TA1, the transaction object ID "M1" is ranked "1st" and the first score SC1 is "1.0". The transaction object ID "M2" is ranked "2nd place" and the first score SC1 is "0.94". The transaction object ID "M3" is ranked "3rd" and the first score SC1 is "0.89". The transaction object ID "M4" is ranked "4th place" and the first score SC1 is "0.84". The transaction object ID "M5" is ranked "5th", and the first score SC1 is "0.79".

The ranking "6th place" is the transaction object ID "M6", and the first score SC1 is "0.74". The ranking "7th place" is the transaction object ID "M7", and the first score SC1 is "0.68". The transaction object ID "M8" is ranked "8th place" and the first score SC1 is "0.63". The transaction object ID "M9" is ranked "9th place" and the first score SC1 is "0.58". The ranking "10th place" is the transaction object ID "M10", and the first score SC1 is "0.53".

In this manner, the selection device 100 determines the ranking of the transaction objects by calculating the first score SC1 for the transaction objects acquired from the external device 20. For example, the selection device 100 ranks the transaction objects in descending order of first score SC1. Further, the above calculation process may be executed every predetermined period or every step S3.

Returning to the explanation of FIG. The selection device 100 selects a candidate for transaction based on the first score SC1 (step S5). For example, the selection device 100 selects the top 20 transaction objects with the highest first scores SC1 as transaction object candidates based on the first score SC1. In the example of FIG. 2, the selection device 100 selects transaction objects whose transaction object IDs are M1 to M10 as transaction object candidates.

Subsequently, the selection device 100 provides the terminal device 10 with content including the top six transaction target candidates ranked based on the first score SC1 (step S6). In the example of FIG. 2, the selection device 100 is a content in which six transaction objects "M1" to "M6" from the order of the highest first score SC1 are arranged in a predetermined display frame, Content related to the portal site is provided to the terminal device 10.

Here, content provided to the terminal device 10 will be explained using FIG. 3. FIG. 3 is a diagram showing a specific example of content according to the embodiment. For example, content C1 is content related to a portal site.

Further, the content C1 includes various information. For example, the content C1 includes a search window, information about news, information about content frequently displayed to users, advertising content, and the like.

Here, the content C1 includes an area AR1 as a predetermined display frame. In the example of FIG. 3, six transaction objects to be traded at online shop A, which is an auction service, are arranged in area AR1. In the example of FIG. 3, the transaction object is a smartphone. For example, in area AR1, the upper left row is "M1" with the ranking "1st", the upper middle row is "M2" with the ranking "2nd", the upper right row is "M3" with the ranking "3rd", and the lower left row is the ranking. The transaction targets are arranged in the following order: "M4" for "4th place", "M5" for "5th place" in the middle lower row, "M6" for "6th place" in the lower right row, and so on.

In steps S1 to S6, an example has been described in which content including a transaction object with a high first score is provided. Next, the selection device 100 calculates a second score based on a predetermined formula (formula (1) described below), and provides content that includes a transaction object with a high second score. An example of this will be explained in steps S7 to S11.

As shown in FIG. 1, the selection device 100 receives a request for content provision from the terminal device 10 (step S7). For example, the selection device 100 receives a request to provide a portal site from the terminal device 10 as an example of content.

Further, the selection device 100 acquires the number of display times for the transaction object from the terminal device 10 (step S8). For example, the selection device 100 acquires the number of times content is displayed and the number of times each transaction object is displayed from the terminal device 10. Here, the number of times of display is the number of times the content or information regarding the transaction object is displayed to the user.

Subsequently, the selection device 100 calculates a second score corresponding to the transaction target candidate based on the number of times of display (step S9). For example, the selection device 100 uses a multi-armed bandit algorithm to calculate the second score corresponding to the transaction target candidate. To give a more specific example, the selection device 100 calculates a second score SC2 based on UCB (Upper Confidence Bound) using the following formula (1).

In Equation 1, SC2 _i indicates the second score of transaction object i, and SC1 _i indicates the first score of transaction object i. Further, in formula (1), T indicates the number of times the content is displayed, N _T (i) indicates the number of times the transaction object i is displayed at the T-th time, and C indicates the initial number of times the transaction object i is displayed.

According to Equation (1), the selection device 100 preferentially selects the transaction object that has been displayed less often, so that the value of the second score becomes larger. Here, using FIG. 4, a specific example of the ranking based on the second score SC2 of the candidate for transaction will be shown. FIG. 4 is a diagram illustrating a specific example of the ranking of transaction target candidates according to the embodiment.

In the example of FIG. 4, a table TA2 is shown that associates the ranking of transaction object candidates, transaction object IDs, first scores SC1, second scores SC2, and display counts. Here, it is assumed that table TA2 shows the first score SC1, second score SC2, and display frequency for ten transaction target candidates.

According to Table TA2, the ranking "1st place" is the transaction target ID "M1", the first score SC1 is "1.0", the second score SC2 is "2.39", and the number of display times is "M1". It is "1.0". The ranking "2nd place" is the transaction target ID "M2", the first score SC1 is "0.94", the second score SC2 is "2.34", and the number of display is "1.0" It is. The ranking "3rd" is the transaction target ID "M3", the first score SC1 is "0.89", the second score SC2 is "2.30", and the number of display is "1.0" It is. The ranking "4th" is the transaction target ID "M4", the first score SC1 is "0.84", the second score SC2 is "2.24", and the number of display is "1.0" It is. The ranking “5th place” is the transaction target ID “M5”, the first score SC1 is “0.79”, the second score SC2 is “2.18”, and the number of display is “1.0” It is.

The ranking "6th place" is the transaction target ID "M6", the first score SC1 is "0.74", the second score SC2 is "2.13", and the number of display is "1.0" It is. The ranking “7th place” is the transaction target ID “M7”, the first score SC1 is “0.68”, the second score SC2 is “2.66”, and the number of display is “0” . The ranking "8th" is the transaction target ID "M8", the first score SC1 is "0.63", the second score SC2 is "2.60", and the number of display is "0" . The ranking "9th place" is the transaction target ID "M9", the first score SC1 is "0.58", the second score SC2 is "2.55", and the number of display times is "0". . The ranking “10th place” is the transaction target ID “M10”, the first score SC1 is “0.53”, the second score SC2 is “2.50”, and the number of display is “0” .

In this way, the selection device 100 calculates the second score SC2 for the transaction target candidate based on the first score SC1 and the number of times of display using the formula (1). Thereby, the selection device 100 determines the ranking of the transaction target candidates based on the second score SC2. For example, the selection device 100 ranks the transaction target candidates in descending order of second score SC2. Further, the above calculation process may be executed every predetermined period or every step S9.

Here, Table TA3 is shown with the ranking changed from Table TA2. Table TA3 is a table that associates the rankings of transaction object candidates, transaction object IDs, first scores SC1, second scores SC2, and display counts.

According to Table TA3, the ranking "1st place" is the transaction target ID "M7", the first score SC1 is "0.68", the second score SC2 is "2.66", and the number of display times is "M7". It is "1.0". The ranking "2nd" is the transaction target ID "M8", the first score SC1 is "0.63", the second score SC2 is "2.60", and the number of display is "1.0" It is. The ranking "3rd" is the transaction target ID "M9", the first score SC1 is "0.58", the second score SC2 is "2.55", and the number of display is "1.0" It is. The ranking "4th" is the transaction target ID "M10", the first score SC1 is "0.53", the second score SC2 is "2.50", and the number of display is "1.0" It is. The ranking "5th" is the transaction target ID "M11", the first score SC1 is "0.48", the second score SC2 is "2.45", and the number of display is "1.0" It is.

The ranking "6th place" is the transaction target ID "M1", the first score SC1 is "1.0", the second score SC2 is "2.39", and the number of display is "2.0" It is. The ranking "7th place" is the transaction target ID "M12", the first score SC1 is "0.43", the second score SC2 is "2.37", and the number of display is "0". . The ranking "8th" is the transaction target ID "M13", the first score SC1 is "0.38", the second score SC2 is "2.36", and the number of display is "0" . The ranking "9th place" is the transaction target ID "M14", the first score SC1 is "0.33", the second score SC2 is "2.35", and the number of display times is "0". . The ranking "10th" is the transaction target ID "M2", the first score SC1 is "0.94", the second score SC2 is "2.34", and the number of display is "1.0" It is. Note that the number of times of display shown in table TA3 is calculated by adding 1 to the number of times of display shown in table TA2, assuming that "1st place" to "6th place" are displayed.

In this way, the selection device 100 uses the multi-arm banded algorithm to select a transaction target from among the transaction target candidates each time the user accesses the content, so that the transaction target with a small number of display times is preferentially displayed. Select again. Thereby, the selection device 100 can appropriately select a transaction target that is highly appealing to the user. In addition, the selection device 100 appropriately selects a transaction object that is not displayed to the user from among the transaction objects that have a high appeal effect to the user, so that the user does not lose an opportunity to purchase the transaction object. can be reduced.

Returning to the explanation of FIG. The selection device 100 selects a transaction target based on the second score (step S10). For example, the selection device 100 selects the top six transaction targets with the highest second scores based on the second scores. In the example of FIG. 4, the selection device 100 selects transaction objects whose transaction object IDs are "M7," "M8," "M9," "M10," "M11," and "M1."

Subsequently, the selection device 100 provides content including the transaction object (step S11). In the example of FIG. 4, the selection device 100 selects six transaction targets "M7", "M8", "M9", "M10", "M11", and "M1" from the order of highest second score. The terminal device 10 is provided with content related to a portal site whose target is arranged in a predetermined display frame.

Then, the selection device 100 executes a selection process of selecting a transaction target based on the number of times the content provided in step S11 is displayed. That is, after executing step S11, the selection device 100 repeats the processing shown in steps S7 to S11.

Note that in the above example, an example was explained in which the selection device 100 provides content including a transaction object with a high first score in steps S1 to S6, but the present invention is not limited to the above example. For example, the selection device 100 may also calculate the second score based on formula (1) in steps S1 to S6, and provide content that includes a transaction object with a high second score. In this case, T shown by formula (1) is 0 and N _T (i) is 0. At this time, the selection device 100 calculates (first score + constant) as the second score. In this way, the selection device 100 may provide content including a transaction object with a high second score.

Conventionally, there have been cases in which it has not always been possible to appropriately select information other than content. Therefore, when a user displays predetermined content, the selection device 100 according to the embodiment selects a transaction target displayed in the predetermined content from a transaction target candidate selected based on user information about the user. The conventional problem is solved by selecting one of them based on the number of times a predetermined content has been displayed.

In this manner, the selection device 100 can propose information for each transaction target using the multi-armed banded algorithm according to the expected value distribution selected by the user. Furthermore, the selection device 100 can provide information on products that the user is likely to desire but do not know about, thereby preventing opportunity loss. For example, even if the desired product on the auction site is sold by another user, the selection device 100 can provide an alternative, such as providing information on other products that are the same or similar. Moreover, since the selection device 100 can reduce opportunity losses such as a user not being able to purchase a transaction object, it is possible to appropriately provide information on the transaction object to be provided to the user.

[2. Selection system configuration]
As shown in FIG. 5, the selection system 1 includes a terminal device 10, an external device 20, and a selection device 100. The terminal device 10, the external device 20, and the selection device 100 are connected via a network N so that they can communicate by wire or wirelessly. Note that the selection system 1 shown in FIG. 5 may include a plurality of terminal devices 10, a plurality of external devices 20, and a plurality of selection devices 100.

The terminal device 10 according to the embodiment is an information processing device used by a user who accesses content such as a web page displayed on a browser or content for an application. For example, the terminal device 10 is a desktop PC (Personal Computer), a notebook PC, a tablet terminal, a mobile phone, a PDA (Personal Digital Assistant), or the like. Note that the terminal device 10 is not limited to the above example, and may be, for example, a smart watch or a wearable device.

The external device 20 according to the embodiment is an information processing device that provides various information, and is realized by, for example, a server device, a cloud system, or the like. For example, the external device 20 provides user information as various information. In addition, as another example, the external device 20 may be used for Internet shopping, an electronic shopping mall, a flea market site, an auction site, a travel or restaurant reservation site, a credit card contract site, a financial product providing site, etc. Provide information regarding transaction objects.

The selection device 100 according to the embodiment is an information processing device that can communicate with various devices via a predetermined network N such as the Internet, and is realized by, for example, a server device or a cloud system. For example, the selection device 100 is communicably connected to various other devices via the network N.

[3. Configuration of selection device]
An example of the functional configuration of the selection device 100 described above will be described below. FIG. 5 is a diagram showing a configuration example of the selection device 100 according to the embodiment. As shown in FIG. 5, the selection device 100 includes a communication section 110, a storage section 120, and a control section 130.

(About communication department 110)
The communication unit 110 is realized by, for example, a NIC (Network Interface Card). The communication unit 110 is connected to the network N by wire or wirelessly, and transmits and receives information to and from various other devices.

(About storage unit 120)
The storage unit 120 is realized by, for example, a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disk. Furthermore, the storage unit 120 includes a user information storage unit 121 , a transaction target information storage unit 122 , and a content storage unit 123 .

(About user information storage unit 121)
The user information storage unit 121 stores various types of user information regarding users. Here, FIG. 6 shows an example of the user information storage unit 121 according to the embodiment. In the example shown in FIG. 6, the user information storage unit 121 has items such as "user ID" and "user information." For example, "user information" includes items such as "attribute information,""purchasehistory," and "display history."

"User ID" is an identifier that identifies a user. "Attribute information" is attribute information regarding the attributes of the user associated with the "user ID." For example, the attribute information includes the user's age, gender, telephone number, address, etc.

"Purchase history" is a history of purchases made by the user associated with the "user ID." For example, the purchase history includes information regarding transaction objects purchased by the user, the type of transaction object, the number of times the transaction object was purchased, the date and time when the transaction object was purchased, and the like.

The "display history" is a history of contents displayed by the user associated with the "user ID." For example, the display history includes information regarding content displayed by the user, the type of content, the number of times the content was displayed, the date and time the content was displayed, and the like.

For example, in FIG. 6, "U1" identified by the user ID has attribute information "CH1", purchase history "PH1", and display history "WA1". In the example shown in Figure 6, the attribute information, etc. is expressed using abstract codes such as "CH1", but the attribute information, etc. can be expressed by specific numerical values, specific character strings, or various information. It may be a file format that includes.

(About transaction target information storage unit 122)
The transaction object information storage unit 122 stores various transaction object information regarding transaction objects. Here, FIG. 7 shows an example of the transaction target information storage unit 122 according to the embodiment. In the example shown in FIG. 7, the transaction object information storage unit 122 includes items such as "transaction object ID", "transaction object", "first score", "second score", and "number of times of display".

“Transaction object ID” is an identifier that identifies a transaction object. The “transaction object” is information regarding the transaction object associated with the “transaction object ID”. The "first score" is the first score associated with the "transaction object ID."

“Second score” is a second score associated with “transaction object ID”. The "number of times of display" is the number of times information regarding the transaction object associated with the "transaction object ID" has been displayed to the user.

For example, in FIG. 7, for "M1" identified by the transaction object ID, the transaction object is "MA1", the first score is "1.0", and the second score is "2.4". , the display count is "1.0". In the example shown in FIG. 7, the transaction object is expressed by an abstract code such as "MA1," but the transaction object may be in a file format containing various information regarding the transaction object.

(About content storage unit 123)
The content storage unit 123 stores information regarding content. Here, FIG. 8 shows an example of the content storage unit 123 according to the embodiment. In the example shown in FIG. 8, the content storage unit 123 has items such as "content ID" and "content".

"Content ID" is an identifier that identifies content. “Content” is information regarding the content associated with “Content ID”. Specifically, the content may indicate information regarding the contents of the content.

For example, the content is content related to a portal site. Furthermore, as another example, the content may be content related to a news site, an auction site, a weather forecast site, a shopping site, a finance (stock price) site, or the like.

Further, the content may be content related to a route search site, a map providing site, a travel site, a restaurant introduction site, a web blog site, an SNS (Social Networking Service) site, or the like.

For example, in FIG. 8, the content of "C1" identified by the content ID is "CO1". Note that in the example shown in Figure 8, the content, etc. is expressed by an abstract code such as "CO1", but the content, etc. is a file containing specific numerical values, specific character strings, and various information. It may be a format, etc.

(About the control unit 130)
The control unit 130 is a controller, and for example, uses a CPU (Central Processing Unit), an MPU (Micro Processing Unit), etc. to control various programs (an example of a selection program) stored in a storage device inside the selection device 100. This is realized by executing using RAM as a work area. Further, the control unit 130 is a controller, and is realized by, for example, an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

As shown in FIG. 5, the control unit 130 includes an acquisition unit 131, a reception unit 132, a calculation unit 133, a first selection unit 134, a second selection unit 135, and a provision unit 136. Realize or execute the information processing functions and actions described in . Note that the internal configuration of the control unit 130 is not limited to the configuration shown in FIG. 5, and may be any other configuration as long as it performs information processing to be described later. Further, the connection relationship between the respective processing units included in the control unit 130 is not limited to the connection relationship shown in FIG. 5, and may be other connection relationships.

(About the acquisition unit 131)
The acquisition unit 131 acquires various information. Specifically, the acquisition unit 131 acquires user information regarding the user from the external device 20. In the example of FIG. 1, the acquisition unit 131 collects, as user information, user attribute information, purchase history on Internet shopping, online shopping malls, auction sites, flea market sites, etc., content display history, etc. from an external device. Obtained from 20. The acquisition unit 131 then stores this user information in the user information storage unit 121. For example, the acquisition unit 131 acquires user information from the external device 20 at predetermined intervals, and updates the user information stored in the user information storage unit 121.

Further, the acquisition unit 131 acquires information regarding the transaction object from the external device 20. For example, the acquisition unit 131 acquires information regarding a transaction object traded at an auction site or a flea market site from the external device 20. To give a more specific example, the acquisition unit 131 acquires from the external device 20 information regarding a smartphone that is being bid on at an auction site. Then, the acquisition unit 131 stores information regarding the transaction target in the transaction target information storage unit 122. For example, the acquisition unit 131 acquires information regarding the transaction object from the external device 20 every predetermined period, and updates the information regarding the transaction object stored in the transaction object information storage unit 122.

The acquisition unit 131 also acquires the number of times the transaction object is displayed from the terminal device 10 . For example, the acquisition unit 131 acquires the number of times content is displayed and the number of times each transaction object is displayed from the terminal device 10. Further, the acquisition unit 131 acquires the content as well as control information provided in advance to acquire the display count. Then, the acquisition unit 131 stores the number of times of display in the transaction target information storage unit 122.

(About the reception section 132)
The reception unit 132 receives various requests. Specifically, the receiving unit 132 receives a request for content provision from the terminal device 10. For example, the receiving unit 132 receives a request for providing a portal site from the terminal device 10 as an example of content.

(About the calculation unit 133)
The calculation unit 133 calculates various scores. Specifically, the calculation unit 133 calculates the first score SC1 of the transaction object acquired by the acquisition unit 131. For example, the calculation unit 133 uses conventional technology such as machine learning to create a learning model that outputs a first score SC1 indicating the possibility of purchasing the transaction object by inputting the transaction object and user information. generate.

For example, the calculation unit 133 generates a learning model that has learned the relationship between the transaction object, user information, and information regarding whether the transaction object was purchased by the user in the past. Thereby, the calculation unit 133 calculates the first score SC1 by inputting the transaction object and the user information into the learning model. Then, the calculation unit 133 stores the first score SC1 in the transaction target information storage unit 122.

Further, the acquisition unit 131 acquires, as learning data, information indicating various types of user information and information on a transaction target selected by the user. Then, when the user information and the information on the transaction object selected by the user are input, the calculation unit 133 inputs the user information and the information on the transaction object not selected by the user. The learning model is trained to output a higher score than when In addition, various other techniques can be employed to quantify the extent to which a user shows interest in a transaction object.

For example, the calculation unit 133 uses deep learning to generate a learning model using a method such as backpropagation. For example, the calculation unit 133 uses a learning process to adjust the weight values that are taken into consideration when transmitting values between nodes. In this way, the calculation unit 133 generates a learning model through a process such as backpropagation that corrects the weight values so that the error between the output of the learning model and the correct answer corresponding to the input is reduced. For example, the calculation unit 133 generates a learning model by performing processing such as backpropagation to minimize a predetermined loss function. Thereby, the calculation unit 133 can perform the learning model generation process. Note that the calculation method is not limited to the above example, and various calculation methods can be adopted based on other machine learning.

In this manner, the calculation unit 133 determines the ranking of the transaction objects by calculating the first score SC1 for the transaction objects acquired from the external device 20. For example, the calculation unit 133 ranks the transaction objects in descending order of the first score SC1. Further, the above calculation process may be executed every predetermined period. Further, the above calculation process may be executed, for example, every step S3 shown in the example of FIG.

Further, the calculation unit 133 calculates a second score corresponding to the transaction target candidate based on the number of times of display. For example, the calculation unit 133 uses a multi-armed bandit algorithm to calculate the second score corresponding to the transaction target candidate. To give a more specific example, the calculation unit 133 calculates the second score SC2 based on the UCB using formula (1). Then, the calculation unit 133 stores this second score SC2 in the transaction target information storage unit 122.

In this way, the calculation unit 133 preferentially selects the transaction object that has been displayed less often, so that the value of the second score becomes larger. Thereby, the calculation unit 133 determines the ranking of the transaction target candidates based on the second score SC2. For example, the calculation unit 133 ranks the transaction target candidates in descending order of second score SC2. Further, the above calculation process may be executed every predetermined period. Further, the above calculation process may be executed, for example, every step S9 shown in the example of FIG.

In addition, the calculation unit 133 selects a transaction target from among the transaction target candidates using a multi-arm banded algorithm each time the user accesses the content so that the transaction target with a small number of display times is preferentially displayed. Try again. Thereby, the calculation unit 133 can appropriately select a transaction target that is highly appealing to users. In addition, the calculation unit 133 appropriately selects a transaction object that is not displayed to the user from among the transaction objects that have a high appeal effect to the user, so that the calculation unit 133 reduces the opportunity loss of the user to purchase the transaction object. can be reduced.

(About the first selection section 134)
The first selection unit 134 selects transaction object candidates from among the transaction objects based on the user information. For example, the first selection unit 134 selects the top 20 transaction objects with the highest first scores SC1 as transaction object candidates based on the first score SC1. In the example of FIG. 2, the first selection unit 134 selects transaction objects whose transaction object IDs are M1 to M10 as transaction object candidates.

(About the second selection section 135)
When the user displays the predetermined content, the second selection unit 135 selects the transaction target displayed in the predetermined content from among the transaction target candidates selected by the first selection unit 134. Select based on number of impressions shown. For example, the second selection unit 135 selects the top six trading targets with the highest second scores from the trading target information storage unit 122 based on the second score. In the example of FIG. 4, the second selection unit 135 selects transaction objects whose transaction object IDs are "M7,""M8,""M9,""M10,""M11," and "M1."

(About the providing unit 136)
The providing unit 136 provides various information. For example, the providing unit 136 refers to the content storage unit 123 and provides the terminal device 10 with content including the top six transaction target candidates ranked based on the first score SC1. In the example of FIG. 2, the providing unit 136 provides content in which six transaction objects "M1" to "M6" from the highest rank of first score SC1 are arranged in a predetermined display frame, Content related to the portal site is provided to the terminal device 10.

Further, the providing unit 136 refers to the content storage unit 123 and provides content including transaction objects ranked based on the second score SC2. In the example of FIG. 4, the providing unit 136 performs transactions for six transaction targets "M7", "M8", "M9", "M10", "M11", and "M1" from the order of the highest second scores. The terminal device 10 is provided with content related to a portal site whose target is arranged in a predetermined display frame.

[4. Processing procedure]
Next, the procedure of the selection process executed by the selection device 100 according to the embodiment will be described using FIG. 9. FIG. 9 is a flowchart illustrating an example of the flow of selection processing executed by the selection device 100 according to the embodiment.

As shown in FIG. 9, the acquisition unit 131 acquires user information from the external device 20 (step S101). The acquisition unit 131 also acquires information regarding the transaction object from the external device 20 (step S102).

Then, the receiving unit 132 receives a request for providing content (step S103). Specifically, if the receiving unit 132 has not received a request for content provision (step S103; No), the process returns to before step S103.

On the other hand, when the reception unit 132 accepts the request for content provision (step S103; Yes), the first selection unit 134 selects users from among the transaction targets based on the user information acquired by the acquisition unit 131. A candidate for transaction to be provided is selected (step S104).

Then, when the user displays the predetermined content, the second selection unit 135 selects the transaction target displayed in the predetermined content from among the transaction target candidates selected by the first selection unit 134. The selection is made based on the number of times the content has been displayed (step S105).

Then, the providing unit 136 provides the terminal device 10 with content including a predetermined number of transaction targets selected by the first selecting unit 134. Specifically, the providing unit 136 provides the terminal device 10 with content in which a predetermined number of transaction objects are arranged in a predetermined display frame (step S106).

Note that in the above flowchart, the reception unit 132 performs step S103 after the acquisition unit 131 executes step S101 and step S102. However, after the reception unit 132 executes step S103, , the acquisition unit 131 may execute step S101 and step S102.

[5. Modified example]
The selection device 100 described above may be implemented in various different forms other than the above embodiment. Therefore, other embodiments of the selection device 100 will be described below.

[5-1. user〕
Although the above embodiment has been described using the example of a user who is registered as a member of a predetermined service, the present invention is not limited to the above example. For example, the user may be any user, such as a user who is not registered as a member of a predetermined service.

[5-2. Transaction target]
In the above embodiment, the transaction object traded on an auction site or a flea market site has been described as an example, but the present invention is not limited to the above example. For example, the transaction object may be any transaction object. For example, the transaction object may be a transaction object whose quantity is limited. Further, the transaction object may be a transaction object that, when purchased by a first user, cannot be purchased by a second user who is a user different from the first user. Further, the transaction object may be a transaction object traded through Internet shopping, an electronic shopping mall, or the like.

[5-3. Select trading targets for each second period that is shorter than the first period]
In the above embodiment, when the user displays the predetermined content, the selection device 100 selects the transaction object displayed in the predetermined content from among the transaction object candidates based on the number of times the predetermined content has been displayed. Although an example of the selection process for selecting based on the above-mentioned information has been described, it is not limited to the above example.

For example, the selection device 100 may select a transaction object from among the transaction object candidates in every second period that is shorter than the first period indicating the time interval at which user information is acquired. The first period here corresponds to a predetermined period during which user information is acquired from the external device 20, as shown in the above example. Thereby, the selection device 100 can select an appropriate transaction target every second period before the user information is updated.

In this way, the selection device 100 according to the embodiment selects a transaction object from among the transaction object candidates every second period that is shorter than the first period indicating the time interval at which user information is acquired. It is possible to appropriately select the information to be traded to be provided to the user.

[5-4. Providing information regarding the transaction object]
Although the above embodiment has been described using an example of the providing process in which the selection device 100 provides content to the terminal device 10, the present invention is not limited to the above example. For example, an external server may provide content to the terminal device 10. In this case, the selection device 100 may provide information regarding the transaction object to the external server.

For example, the external server provides content related to a portal site as an example of the content. In this case, the selection device 100 may provide the external server with an image of the transaction object, information regarding the layout when placed in the content, etc. as information regarding the transaction object.

[5-5. Transaction target information storage unit 122]
Although the above embodiment has been described using an example in which the transaction target information storage unit 122 stores the number of times of display, the present invention is not limited to the above example. For example, the transaction object information storage unit 122 does not need to store the number of times of display. In this case, the user information storage unit 121 may store the number of times of display for each user. Note that various methods can be used to manage the number of displays. For example, the user information storage unit 121 may manage the number of times of display for each user as a display history.

[5-6. program〕
Further, the selection device 100 according to the embodiment described above is realized, for example, by a computer 1000 having a configuration as shown in FIG. FIG. 10 is a diagram showing an example of a hardware configuration. The computer 1000 is connected to an output device 1010 and an input device 1020, and has an arithmetic device 1030, a primary storage device 1040, a secondary storage device 1050, an output IF (Interface) 1060, an input IF 1070, and a network IF 1080 connected by a bus 1090. has.

The arithmetic unit 1030 operates based on programs stored in the primary storage device 1040 and the secondary storage device 1050, programs read from the input device 1020, and performs various processes. The primary storage device 1040 is a memory device such as a RAM that temporarily stores data used by the arithmetic unit 1030 for various calculations. Further, the secondary storage device 1050 is a storage device in which data used by the arithmetic unit 1030 for various calculations and various databases are registered, and includes ROM (Read Only Memory), HDD (Hard Disk Drive), flash memory, etc. This is realized by

The output IF 1060 is an interface for transmitting information to be output to an output device 1010 that outputs various information such as a monitor or a printer, and is, for example, a USB (Universal Serial Bus), a DVI (Digital Visual Interface), This is realized using a connector compliant with standards such as HDMI (registered trademark) (High Definition Multimedia Interface). Further, the input IF 1070 is an interface for receiving information from various input devices 1020 such as a mouse, a keyboard, and a scanner, and is realized by, for example, a USB or the like.

Note that the input device 1020 is, for example, an optical recording medium such as a CD (Compact Disc), a DVD (Digital Versatile Disc), or a PD (Phase change rewritable disk), a magneto-optical recording medium such as an MO (Magneto-Optical disk), or a tape. It may be a device that reads information from a medium, a magnetic recording medium, a semiconductor memory, or the like. Further, the input device 1020 may be an external storage medium such as a USB memory.

Network IF 1080 receives data from other devices via network N and sends it to computing device 1030, and also sends data generated by computing device 1030 to other devices via network N.

Arithmetic device 1030 controls output device 1010 and input device 1020 via output IF 1060 and input IF 1070. For example, the arithmetic device 1030 loads a program from the input device 1020 or the secondary storage device 1050 onto the primary storage device 1040, and executes the loaded program.

For example, when the computer 1000 functions as the selection device 100, the arithmetic unit 1030 of the computer 1000 realizes the functions of the control unit 130 by executing a program loaded onto the primary storage device 1040.

[5-7. others〕
Furthermore, among the processes described in the above embodiments and modified examples, all or part of the processes described as being performed automatically can be performed manually, or may be described as being performed manually. All or part of this processing can also be performed automatically using known methods. In addition, information including the processing procedures, specific names, and various data and parameters shown in the above documents and drawings may be changed arbitrarily, unless otherwise specified. For example, the various information shown in each figure is not limited to the illustrated information.

Furthermore, each component of each device shown in the drawings is functionally conceptual, and does not necessarily need to be physically configured as shown in the drawings. In other words, the specific form of distributing and integrating each device is not limited to what is shown in the diagram, and all or part of the devices can be functionally or physically distributed or integrated in arbitrary units depending on various loads and usage conditions. Can be integrated and configured.

Furthermore, the above-described embodiments and modifications can be combined as appropriate within a range that does not conflict with the processing contents.

Further, the above-mentioned "section, module, unit" can be read as "means", "circuit", etc. For example, the selection section can be replaced with selection means or selection circuit.

[6. effect〕
As described above, the selection device 100 according to the embodiment includes the first selection section 134 and the second selection section 135. The first selection unit 134 selects a transaction object candidate, which is a candidate for a transaction object to be provided to the user, from among the transaction objects based on user information regarding the user. When the user displays the predetermined content, the second selection unit 135 selects the transaction target displayed in the predetermined content from among the transaction target candidates selected by the first selection unit 134. Select based on number of impressions shown.

Thereby, when the user displays the predetermined content, the selection device 100 according to the embodiment selects the transaction object displayed in the predetermined content as the transaction object candidate selected based on the user information about the user. Since the selection is made based on the number of times a predetermined content has been displayed, it is possible to appropriately select the information to be traded to be provided to the user.

In the selection device 100 according to the embodiment, the second selection unit 135 selects a relationship between the number of times a predetermined content is displayed and the number of times a trade object displayed in the predetermined content is displayed from among the transaction object candidates. Select trading targets based on

Thereby, the selection device 100 according to the embodiment selects a transaction object from among the transaction object candidates based on the relationship between the number of times a predetermined content is displayed and the number of times a transaction object displayed in the predetermined content is displayed. Therefore, it is possible to appropriately select the information on the transaction target to be provided to the user.

In addition, in the selection device 100 according to the embodiment, the second selection unit 135 is configured to perform a transaction when the first display number of the first transaction object displayed in the predetermined content is smaller than the second display number of the second transaction object. A first transaction target is selected from among the target candidates.

Thereby, the selection device 100 according to the embodiment selects from among the transaction object candidates when the first display number of the first transaction object displayed in the predetermined content is smaller than the second display number of the second transaction object. Since the first transaction object is selected, the information on the transaction object to be provided to the user can be appropriately selected.

In addition, in the selection device 100 according to the embodiment, the second selection unit 135 selects a transaction target from among the transaction target candidates for each second period shorter than the first period indicating the time interval at which user information is acquired. Select.

As a result, the selection device 100 according to the embodiment selects a transaction object from among the transaction object candidates in each second period that is shorter than the first period indicating the time interval at which user information is acquired. It is possible to appropriately select the information to be provided to the transaction target.

Further, in the selection device 100 according to the embodiment, the second selection unit 135 selects a transaction target from among the transaction target candidates using a multi-armed bandit algorithm.

Thereby, the selection device 100 according to the embodiment appropriately selects the information on the transaction object to be provided to the user in order to select the transaction object from among the transaction object candidates using the multi-armed bandit algorithm. be able to.

Furthermore, in the selection device 100 according to the embodiment, the first selection unit 134 selects transaction object candidates from among the transaction objects for each predetermined period based on the user information.

Thereby, the selection device 100 according to the embodiment selects transaction target candidates from among the transaction targets for each predetermined period based on the user information, and the information on the transaction targets provided to the user. can be selected appropriately.

The selection device 100 according to the embodiment further includes a providing unit 136 that provides content including a predetermined number of transaction targets selected by the second selection unit 135.

Thereby, the selection device 100 according to the embodiment provides content including a predetermined number of selected transaction objects, and therefore can provide content including transaction objects that are highly appealing to users.

Furthermore, in the selection device 100 according to the embodiment, the providing unit 136 provides content in which a predetermined number of transaction targets are arranged in a predetermined display frame.

Thereby, the selection device 100 according to the embodiment arranges transaction objects that are highly appealing to users in the predetermined display frames in order to provide content in which a predetermined number of transaction objects are arranged in the predetermined display frames. content can be provided.

Furthermore, in the selection device 100 according to the embodiment, the transaction object is a transaction object whose quantity is limited.

In addition, in the selection device 100 according to the embodiment, the transaction object is a transaction object that, when purchased by the first user, cannot be purchased by a second user who is a user different from the first user. It is.

Further, in the selection device 100 according to the embodiment, the transaction object is a transaction object traded at an auction or a flea market.

In this way, the selection device 100 according to the embodiment can provide content including various transaction objects that are highly appealing to users.

Some of the embodiments of the present application have been described above in detail based on the drawings, but these are merely examples, and various modifications and variations can be made based on the knowledge of those skilled in the art, including the embodiments described in the disclosure section of the invention. It is possible to carry out the invention in other forms with modifications.

N Network 1 Selection system 10 Terminal device 20 External device 100 Selection device 110 Communication unit 120 Storage unit 121 User information storage unit 122 Transaction target information storage unit 123 Content storage unit 130 Control unit 131 Acquisition unit 132 Reception unit 133 Calculation unit 134 1 selection section 135 2nd selection section 136 provision section

Claims (13)

Using a learning model that outputs a score indicating the possibility that the specified user will purchase the specified transaction object by inputting the attributes of a specified user and the specified transaction object from among the transaction objects. a first selection unit that selects a transaction target whose calculated score exceeds a predetermined threshold as a transaction target candidate that is a candidate for a transaction target to provide information to the predetermined user;
When providing a predetermined content provided to each user, in which information regarding a plurality of transaction objects is arranged side by side, to the predetermined user, the first selection unit Based on the history of the number of times the predetermined content has been displayed from among the transaction target candidates selected based on the score, and the number of times information has been arranged and displayed in the predetermined content for each of the transaction target candidates. , and a second selection unit that selects a transaction target candidate whose information is to be placed in a predetermined content provided to the predetermined user. The second selection section includes:
Based on the history of the number of times the predetermined content has been displayed and the number of times information has been arranged and displayed in the predetermined content for each candidate for transaction, information on each candidate for transaction is provided to the predetermined user. calculate a score to be placed in the predetermined content provided to the user, and based on the calculated score, select a transaction object from the transaction object candidates for placing information in the predetermined content provided to the predetermined user. The selection device according to claim 1, characterized in that: The second selection section includes:
The first display count, which is the number of times information regarding the first transaction object that is the transaction object candidate is displayed in the predetermined content, is the number of times information regarding the second transaction object that is the transaction object candidate is displayed in the predetermined content. If the number of display times is smaller than a second display number, the first transaction object is selected as a transaction object candidate for placing information in a predetermined content provided to the predetermined user. 3. The selection device according to 1 or 2. The second selection section includes:
The transaction object is selected from among the transaction object candidates every second period that is shorter than a first period indicating a time interval at which attributes of the predetermined user are acquired. 3. The selection device according to any one of 3. The second selection section includes:
Each time the predetermined content is provided to any user, based on the history of the number of times the predetermined content has been displayed and the number of times information has been arranged and displayed in the predetermined content for each candidate for transaction. The selection device according to claim 2, wherein a transaction target for which information is to be placed in the predetermined content is reselected from the transaction target candidates using a multi-armed bandit algorithm. The first selection section includes:
Based on the score calculated using a learning model that outputs the score by inputting the attributes of the predetermined user and the predetermined transaction target based on information acquired every predetermined period, The selection device according to any one of claims 1 to 5, characterized in that the transaction target candidate is selected from among the candidates. The selection device according to any one of claims 1 to 6, further comprising a providing unit that provides content including a predetermined number of transaction objects selected by the second selection unit. The provision department is
The selection device according to claim 7, wherein content is provided in which the predetermined number of transaction objects are arranged in a predetermined display frame. The said transaction object is
The selection device according to any one of claims 1 to 8, characterized in that the selection device is a transaction object whose quantity is limited. The said transaction object is
Any one of claims 1 to 9, characterized in that, when purchased by a first user, the transaction object cannot be purchased by a second user who is a user different from the first user. The selection device according to any one of the above. The said transaction object is
The selection device according to any one of claims 1 to 10, characterized in that the selection device is a transaction object traded at an auction or a flea market. A selection method performed by a computer, comprising:
Using a learning model that outputs a score indicating the possibility that the specified user will purchase the specified transaction object by inputting the attributes of a specified user and the specified transaction object from among the transaction objects. a first selection step of selecting a transaction target whose calculated score exceeds a predetermined threshold as a transaction target candidate that is a candidate for a transaction target to provide information to the predetermined user;
When providing a predetermined content provided to each user, in which information regarding a plurality of transaction objects is arranged side by side, to the predetermined user, the first selection step Based on the history of the number of times the predetermined content has been displayed from among the transaction target candidates selected based on the score, and the number of times information has been arranged and displayed in the predetermined content for each of the transaction target candidates. , and a second selection step of selecting transaction target candidates whose information is to be placed in a predetermined content provided to the predetermined user. Using a learning model that outputs a score indicating the possibility that the specified user will purchase the specified transaction object by inputting the attributes of a specified user and the specified transaction object from among the transaction objects. a first selection step of selecting a transaction target whose calculated score exceeds a predetermined threshold as a transaction target candidate that is a candidate for a transaction target to provide information to the predetermined user;
When providing a predetermined content provided to each user, in which information regarding a plurality of transaction objects is arranged side by side, to the predetermined user, the first selection step Based on the history of the number of times the predetermined content has been displayed from among the transaction target candidates selected based on the score, and the number of times information has been arranged and displayed in the predetermined content for each of the transaction target candidates. , a second selection procedure for selecting transaction target candidates whose information is to be arranged in a predetermined content provided to the predetermined user.

Images