JP6243559B2 - Learning device, learning method, and learning program - Google Patents

Description

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

  In recent years, with the rapid spread of the Internet, information provision through the Internet has been actively performed. An example of information provision is distribution of Internet advertisements. For example, in a mobile terminal such as a smartphone, an advertisement such as a banner is displayed together with the content on a display screen on which an application program (hereinafter also referred to as “app”) or a browser displays the content. When the user clicks such an advertisement, for example, an application download site associated with the advertisement is displayed.

  The user may install the application on the mobile terminal through the advertisement as described above. As a technique related to application installation, a technique for transmitting only information related to a specific installed application to a server is known (for example, Patent Document 1). Also, a technique is known in which when a selection operation by a user is performed offline, the selection operation is stored to execute a predetermined process when the user enters an online state (for example, Patent Document 2). ).

JP 2014-167688 A JP2013-257683A

  However, in the above-described conventional technology, matching between the provided information content and the user of the providing destination is not always performed appropriately. Since delivery of advertisements (an example of information content) is expensive, it is desirable for advertisers to preferentially deliver advertisements to users who can easily demonstrate the effects of advertisements. However, although the above-mentioned conventional technology can measure the result of the distributed advertisement, it accurately extracts users who are likely to exert the effect of the advertisement, such as a user who can easily install an application through the advertisement. It ’s difficult.

  The present application has been made in view of the above, and an object thereof is to provide a learning apparatus, a learning method, and a learning program capable of appropriately matching information content to be provided and a user to whom the information is provided. .

  The learning device according to the present application includes an acquisition unit that acquires terminal information that is information related to a terminal device that receives information content, result information indicating whether or not a predetermined action related to predetermined information content has been executed, and the acquisition unit Based on the model generated by the generation unit, the generation unit generating a model showing a tendency to execute a predetermined action related to the predetermined information content by learning the relationship with the terminal information acquired by A specifying unit that specifies a terminal device that provides the predetermined information content, and the acquisition unit is an application that is information about an application installed in the terminal device as the terminal information The information is acquired, and the generation unit generates the model using the application information.

  According to one aspect of the embodiment, there is an effect that matching between the provided information content and the user of the providing destination can be performed appropriately.

FIG. 1 is a diagram illustrating an example of a learning process according to the embodiment. FIG. 2 is a diagram (1) illustrating an example of terminal information according to the embodiment. FIG. 3 is a diagram (2) illustrating an example of terminal information according to the embodiment. FIG. 4 is a diagram illustrating a configuration example of the learning processing system according to the embodiment. FIG. 5 is a diagram illustrating a configuration example of the learning device according to the embodiment. FIG. 6 is a diagram illustrating an example of the advertisement information storage unit according to the embodiment. FIG. 7 is a diagram illustrating an example of an attribute table according to the embodiment. FIG. 8 is a diagram illustrating an example of a device table according to the embodiment. FIG. 9 is a diagram illustrating an example of the application table according to the embodiment. FIG. 10 is a diagram (1) illustrating an example of the setting table according to the embodiment. FIG. 11 is a diagram (2) illustrating an example of the setting table according to the embodiment. FIG. 12 is a diagram illustrating an example of a variable table according to the embodiment. FIG. 13 is a diagram illustrating an example of a model table according to the embodiment. FIG. 14 is a diagram illustrating a configuration example of a user terminal according to the embodiment. FIG. 15 is a flowchart (1) illustrating a processing procedure according to the embodiment. FIG. 16 is a flowchart (2) illustrating a processing procedure according to the embodiment. FIG. 17 is a hardware configuration diagram illustrating an example of a computer that realizes the function of the learning apparatus.

  Hereinafter, a mode for carrying out a learning device, a learning method, and a learning program according to the present application (hereinafter referred to as “embodiment”) will be described in detail with reference to the drawings. Note that the learning device, the learning method, and the learning program according to the present application are not limited to the embodiment. In addition, the embodiments can be appropriately combined within a range that does not contradict processing contents. In the following embodiments, the same portions are denoted by the same reference numerals, and redundant description is omitted.

[1. Example of learning process)
First, an example of learning processing according to the embodiment will be described with reference to FIGS. 1 to 3. FIG. 1 is a diagram illustrating an example of a learning process according to the embodiment. In Figure 1, the learning apparatus 100 according to the present application, the result information indicating whether or not perform the predetermined action on predetermined information content, and the terminal information that the user terminal 10 ₁ as a providing destination of the predetermined information content has An example is shown in which processing for learning the relationship is performed. In the embodiment, an advertisement that is displayed on a web page as information content and that uses a predetermined application as an advertisement target will be described as an example. In this case, the information provider is an advertiser. In the embodiment, an example of an action of installing an application corresponding to a predetermined advertisement in the terminal device will be described as an example of the predetermined action regarding the predetermined information content.

  A learning device 100 illustrated in FIG. 1 is a server device that holds an advertisement submitted by an advertiser. When receiving a request for advertisement distribution from a terminal device operated by the user, the learning device 100 selects an advertisement to be distributed to the terminal device from among the held advertisements. Then, the learning device 100 distributes the selected advertisement to the terminal device.

The user terminal 10 ₁ and 10 ₂ shown in FIG. 1 is an information processing terminal such as a smartphone. In embodiments, the user terminal 10 ₁ is utilized by the user U01. The user terminal 10 ₂ is used by the user U02. The user terminal 10 ₁ and 10 ₂ are obtained content (e.g., web pages) when there is the advertisement display area is to serve ads to be displayed on the advertisement display area, it requests the learning apparatus 100. Then, the user terminals 10 ₁ and 10 ₂ display the acquired advertisement in the advertisement display area. In the following, when it is not necessary to distinguish the user terminals 10 ₁ and 10 ₂ are referred to as "user terminal 10".

  An advertiser terminal 20 shown in FIG. 1 is a terminal device used by an advertiser. For example, the advertiser terminal 20 submits an advertisement to the learning device 100 according to an operation by the advertiser. In the embodiment, an application is associated with an advertisement submitted by an advertiser as an object of promotion. For example, when the user clicks on an advertisement displayed on the user terminal 10, the user terminal 10 installs an application associated with the advertisement. Alternatively, the user terminal 10 transitions the screen display to an application download page or the like associated with the advertisement.

  In FIG. 1, when receiving a request for advertisement distribution from the user terminal 10, the learning apparatus 100 determines an appropriate advertisement to be distributed to the user terminal 10, that is, an advertisement that is assumed to have high advertising effectiveness. For example, when the learning device 100 distributes an advertisement associated with an application, whether the application associated with the advertisement is likely to be installed in the user terminal 10 in the user terminal 10 that is the distribution destination. Determine whether or not. The learning device 100 generates a model for evaluating the possibility that the application is installed in the user terminal 10 for the above determination process. And the learning apparatus 100 specifies the user terminal 10 used as the delivery destination of an advertisement using the produced | generated model. For example, the learning device 100 specifies the distribution-destination user terminal 10 by calculating an index value such as a probability that an application is installed when an advertisement is distributed to the user terminal 10 based on the model. Hereinafter, an example of the learning process performed by the learning apparatus 100 will be described along the flow with reference to FIG. 1.

First, the learning device 100 receives an advertisement related to the application from the advertiser via the advertiser terminal 20 (step S11). The learning device 100 stores the received advertisement in the advertisement information storage unit 121. In addition, the learning device 100 identifies an application associated with the advertisement. Then, the learning apparatus 100 performs model generation processing with the application as a processing target. That is, the learning device 100 generates a model for each application. In other words, when an application and an advertisement are associated with each other on a one-to-one basis, a model is generated for each advertisement. Note that the learning device 100 distributes the advertisement submitted from the advertiser to the user terminal 10 without using the model until the information necessary for generating the model is accumulated (step S12). In the example of FIG. 1, a user terminal 10 which ads were served without the model it is used, and be referred to as user terminal 10 _1.

Subsequently, the learning device 100, advertisements are processed app is correlated with the user terminal 10 ₁ delivered, it acquires the terminal information which is information relating to the user terminal 10 ₁ (step S13). Incidentally, although not shown in FIG. 1, _one user terminal 10 to which ads were served shall present enough considerable number to perform the learning process. For details on terminal information will be described later, the terminal information includes characteristics information indicating each user's identity utilizing user terminal 10 _1. Therefore, the learning apparatus 100, by acquiring the terminal information, it is possible to catch a tendency or has been utilized by the user U01 with what characteristics the user terminal 10 _1. The learning device 100 stores the acquired terminal information in the terminal information storage unit 122.

Further, the learning apparatus 100, the user terminal 10 ₁ ads associated with the application has been delivered to be processed, actually app acquires such result information whether it is installed. That is, the learning apparatus 100, for each user terminal 10 ₁ ad is served, acquires the result information such whether application is installed, and acquires the terminal information from the user terminal 10 _1.

Then, the learning apparatus 100 includes a result information indicating whether or not to install the application that has been processed by learning the relationship between the obtained terminal information from the user terminal 10 ₁ which ads were served, processes Generate a model that shows the tendency to install the targeted app. In other words, the learning device 100 performs machine learning related to the user U01 who tends to install an application (step S14). More specifically, the learning apparatus 100 assumes that the result information indicating whether or not the application to be processed is installed is an objective variable, and is information extracted from the terminal information and indicates the user's characteristics. A model with the feature information to be explained as an explanatory variable is generated. The learning device 100 stores the generated model in the learning model storage unit 126.

Learning apparatus 100 uses the generated model, based on the results of the machine learning, to deliver advertisements for application to the user terminal 10 ₂ (step S15). By using the generated model, the learning apparatus 100 can accurately identify a user who tends to install an application that is a processing target. For example, the learning apparatus 100, when receiving a request for advertisements distributed from the user terminal 10 _2, and acquires the terminal information of the user terminal 10 _2. Then, the learning apparatus 100 inputs the terminal information obtained from the user terminal 10 ₂ is ad delivery destination candidates in the model. Learning apparatus 100 determines based on the information output from the model, whether or not the user terminal 10 ₂ is being utilized by the user U02 likely to install the app. Learning apparatus 100, when the user terminal 10 ₂ is determined to have been utilized by the user U02 likely to install application identifies the user terminal 10 ₂ as the delivery destination ad. Then, the learning apparatus 100 delivers the advertisement to which the application is associated with the user terminal 10 _2.

  Thus, the learning device 100 acquires terminal information that is information regarding the user terminal 10 that receives the advertisement. Further, the learning device 100 learns the relationship between the result information indicating whether or not the application associated with the advertisement has been installed and the acquired terminal information, so that the application associated with the advertisement is performed. Generate a model that shows the tendency to install. Furthermore, the learning device 100 specifies the user terminal 10 that is a delivery destination for delivering the advertisement based on the generated model.

  That is, the learning device 100 generates a model that determines a user who is likely to perform a predetermined action, such as installing an application associated with an advertisement, based on the terminal information. And learning device 100 can distribute an advertisement to a user who tends to be similar to a user who installed an application in the past by specifying a distribution destination of an advertisement based on a model. For this reason, the learning apparatus 100 can increase the possibility that an app associated with the advertisement is installed as compared with a case where the advertisement is distributed without using a model. In other words, the learning device 100 can improve the advertising effect exhibited by the distributed advertisement. In this way, the learning apparatus 100 appropriately matches the distributed advertisement and the distribution destination user.

Incidentally, in FIG. 1, the learning device 100, advertisement showing an example of generating a model by using the terminal information of the user terminal 10 ₁ delivered. In this case, in order to increase the accuracy of the model, it is desirable that the learning apparatus 100 extract a lot of information that accurately characterizes the user U01 from the terminal information. Here, an example of terminal information acquired by the learning device 100 will be described with reference to FIGS. 2 and 3.

  FIG. 2 is a diagram (1) illustrating an example of terminal information according to the embodiment. In FIG. 2, as an example of terminal information, information (hereinafter, referred to as “apparatus information”) held by the apparatus of the user terminal 10 will be described.

  For example, when the learning device 100 receives a request for advertisement distribution from the user terminal 10, the learning device 100 acquires the terminal information of the user terminal 10. The device information included in the terminal information is, for example, the model number 40 set in the user terminal 10. For example, the learning device 100 acquires the information of the model number 40 by acquiring device-specific identification information transmitted from the user terminal 10. Furthermore, the learning device 100 also acquires the feature information 42 extracted from the model number 40 as an example of device information. For example, when the brand name is given to the user terminal 10 by the manufacturer that manufactured the user terminal 10, the learning device 100 extracts the brand name as feature information based on the model number 40. In the example of FIG. 2, it is assumed that the brand name “AAA” is given to the user terminal 10 by the manufacturer. Note that the learning apparatus 100 can perform the above extraction process by referring to a predetermined database in which the correspondence between the model number 40 and the brand name “AAA” is stored.

  Also, the learning device 100 extracts, based on the model number 40, for example, feature information that “336 days” have passed since the user terminal 10 was released. Further, the learning device 100 extracts feature information indicating that the communication carrier is “BBB” based on the model number 40. Further, the learning device 100 extracts feature information indicating that the screen resolution of the user terminal 10 is “1280 × 720” based on the model number 40.

  The learning apparatus 100 uses the model number 40 extracted from the terminal information, the feature information 42, and the like as elements that characterize the user terminal 10. In other words, the learning device 100 extracts the characteristics of the user who uses the user terminal 10 based on the terminal information. This indicates that the model number 40 and the feature information 42 of the user terminal 10 function as elements that characterize the user.

  For example, it is assumed that the brand name “AAA” of the user terminal 10 is a brand generally favored by men and has a sophisticated image. In this case, a user who uses the user terminal 10 to which the brand name “AAA” is assigned is assumed to be a person who likes a sophisticated image. Thus, the learning apparatus 100 can use the brand name of the user terminal 10 as one element that characterizes the person U01. In addition, the learning apparatus 100 treats “the number of days that have passed since the release” as an element of the characterization of whether or not the user using the user terminal 10 prefers a relatively new item. Further, the learning apparatus 100 treats the “communication carrier” as an element of characterization such as whether the user using the user terminal 10 desires a stable communication line or wants a low fee. In addition, the learning apparatus 100 treats “resolution” as an element of characterization such as whether or not the user likes a relatively large screen.

  And the learning apparatus 100 produces | generates the model which reflected the user's characteristic by using the acquired apparatus information as one of the explanatory variables in a model. For example, the learning device 100 learns whether the user has a predetermined relationship with the tendency that the user installs the processing target application and that the brand name of the user terminal 10 used by the user is “AAA”. To do. Similarly, the learning device 100 learns the relationship between the tendency of the user to install the processing target application, the number of days after the user terminal 10 is released, the communication carrier, and the resolution. As described above, the learning device 100 can use information on the device itself of the user terminal 10 as an example of an element characterizing the user of the user terminal 10.

  Further, the learning apparatus 100 can use different information for learning as terminal information. This point will be described with reference to FIG. FIG. 3 is a diagram (2) illustrating an example of terminal information according to the embodiment. In FIG. 3, as an example of terminal information, information related to an application installed in the user terminal 10 (hereinafter referred to as “application information”) will be described. FIG. 3 shows a plurality of applications installed in the user terminal 10.

  In the example illustrated in FIG. 3, the user of the user terminal 10 installs the communication applications 50 and 52 that perform communication with other users in the user terminal 10. In addition, the user installs a traffic application 54, a finance application 56, a shopping application 58 or 60, an exercise application 62 or 64, a rental search application 66, a child care application 68, a strategy game application 70, and a voice actor game application 72. ing.

  Here, the learning apparatus 100 refers to a predetermined database in which an association between an application installed in the user terminal 10 and feature information that is an element characterizing the user is stored. For example, in the predetermined database, it is assumed that the communication applications 50 and 52 are associated with characteristic information such as “lifestyle” and “SNS (Social Networking Service)”. This is because a user who uses an application of a type that mainly communicates with other users, such as communication applications 50 and 52, tends to be interested in elements such as “lifestyle” and “SNS”. Based on the assumption of being a user. Similarly, the traffic system application 54 is associated with feature information such as “transfer guidance”, and the finance application 56 is associated with feature information such as “stock price / exchange”.

  Then, the learning device 100 extracts feature information that is an element that characterizes the user based on the application installed in the user terminal 10. For example, the learning device 100 acquires terminal information that the communication applications 50 and 52 are installed in the user terminal 10. In this case, the learning apparatus 100 characterizes that the user who uses the user terminal 10 has an interest in “lifestyle” or “SNS”.

  Specifically, when the communication application 50 or 52 is installed in the user terminal 10, the learning apparatus 100 uses variables such as “interest_lifestyle” and “interest_SNS” as explanatory variables for explaining the user terminal 10. Is given to the user terminal 10. In addition, when the transportation application 54 and the finance application 56 are installed, the learning apparatus 100 appropriately assigns variables such as “interest_transfer guidance” and “interest_stock price / exchange” to the user terminal 10.

  Note that when an app related to a game is installed in the user terminal 10, the learning device 100 may give a variable different from a variable used in an app other than the game. This is because an application related to a game can be an element that clearly shows the user's characteristics by subdividing the user's characteristics for each type of game.

  For example, when a game application is installed in the user terminal 10, the learning device 100 gives the user terminal 10 a variable classified by the type “game preference” as an explanatory variable for explaining the user terminal 10. Specifically, when the strategic game application 70 is installed in the user terminal 10, the learning device 100 gives a variable such as “game preference_strategy” or “game preference_simulation” to the user terminal 10. In addition, when the voice actor-based game application 72 is installed, the learning device 100 gives variables such as “game preference_voice actor” and “game preference_nurturing” to the user terminal 10. The feature information 80, which is a set of variables extracted from the application installed on the user terminal 10, is generated for each user terminal 10 (that is, for each user).

  As described above, the learning apparatus 100 can generate a model reflecting the user's characteristics by using the acquired application information as an element of an explanatory variable in the model. That is, the learning apparatus 100 learns whether the user has a tendency to install the processing target application and what kind of relationship it has with the application used by the user. For example, the learning apparatus 100 learns the relationship between a user who has installed a processing target application and an application other than the processing target used by the user. Thereby, the learning apparatus 100 can generate a model that can evaluate the possibility that the user using the application is likely to install the application to be processed in the future. As described above, the learning device 100 uses application information of an application installed on the user terminal 10 as an example of an element that characterizes the user of the user terminal 10. Although details will be described later, the learning apparatus 100 can also use the application information for learning as to the total number of applications installed in the user terminal 10, the number of game applications and applications other than games, and the like.

  As described above, the learning device 100 uses the terminal information that can be acquired from the user terminal 10 to generate a model indicating a tendency that the processing target application is installed in the user terminal 10. Hereinafter, the configuration of the learning apparatus 100 that performs such processing and the learning processing system 1 including the learning apparatus 100 will be described in detail.

[2. Learning processing system configuration]
A configuration of the learning processing system 1 including the learning device 100 according to the embodiment will be described with reference to FIG. FIG. 4 is a diagram illustrating a configuration example of the learning processing system 1 according to the embodiment. As illustrated in FIG. 4, the learning processing system 1 according to the embodiment includes a user terminal 10, an advertiser terminal 20, a web server 30, and a learning device 100. These various devices are communicably connected via a network N by wire or wireless. Note that the learning processing system 1 illustrated in FIG. 4 may include a plurality of user terminals 10, a plurality of advertiser terminals 20, and a plurality of web servers 30.

  The user terminal 10 is, for example, an information processing apparatus such as a smartphone, a desktop PC (Personal Computer), a notebook PC, a tablet terminal, a mobile phone, a PDA (Personal Digital Assistant), or a wearable device. It is. The user terminal 10 acquires the web page from the website provided from the web server 30 by accessing the web server 30 according to the operation by the user. Then, the user terminal 10 displays the acquired web page on a display device (for example, a liquid crystal display). In the present specification, the user and the user terminal 10 may be identified with each other. For example, “providing information content to the user” may actually mean “providing information content to the user terminal 10 used by the user”.

  The advertiser terminal 20 is an information processing device used by an advertiser who requests the learning device 100 to distribute an advertisement. The advertiser terminal 20 submits an advertisement related to the application to the learning device 100 according to an operation by the advertiser.

  Note that the advertiser may use the advertiser terminal 20 to request an agency for such submission without submitting the advertisement to the learning device 100. In this case, it is the agency that submits the advertisement to the learning apparatus 100. In the following, the term “advertiser” is a concept that includes not only the advertiser but also the agency, and the term “advertiser terminal” includes the agency device used by the agency as well as the advertiser terminal. It is assumed to be a concept.

  The web server 30 is a server device that provides various web pages when accessed from the user terminal 10. The web server 30 provides various web pages related to, for example, a news site, a weather forecast site, a shopping site, a finance (stock price) site, a route search site, a map providing site, a travel site, a restaurant introduction site, and a web blog.

  As described above, the web page provided by the web server 30 includes an advertisement frame that is a display area for displaying an advertisement. The web page including the advertising space includes an acquisition command for acquiring information content to be displayed in the advertising space. For example, in an HTML (HyperText Markup Language) file that forms a web page, the URL of the learning device 100 is described as an acquisition command. The user terminal 10 that has acquired the web page receives an advertisement from the learning device 100 by accessing a URL described in an HTML file or the like.

  The learning device 100 is a server device that appropriately identifies the user terminal 10 serving as a distribution destination for the advertisement received from the advertiser terminal 20. As described above, the learning device 100 generates a model based on result information indicating whether or not an application is installed in the user terminal 10 and the terminal information of the user terminal 10. And the learning apparatus 100 specifies the user terminal 10 used as a delivery destination using the produced | generated model.

  As described above, the learning device 100 identifies the user terminal 10 or acquires the terminal information of the user terminal 10 when distributing the advertisement. For example, the terminal information of the user terminal 10 can be acquired by including information in a cookie exchanged between the learning device 100 and the web browser or browser application of the user terminal 10. However, the method of acquiring terminal information is not limited to the above. For example, a dedicated program may be set in the user terminal 10 and terminal information may be transmitted to the learning device 100 from the dedicated program. Further, the learning device 100 may acquire the terminal information of the user terminal 10 from the web server 30 that has received access from the user terminal 10.

[3. Configuration of learning device]
Next, the configuration of the learning device 100 according to the embodiment will be described with reference to FIG. FIG. 5 is a diagram illustrating a configuration example of the learning device 100 according to the embodiment. As illustrated in FIG. 5, the learning device 100 includes a communication unit 110, a storage unit 120, and a control unit 130. The learning device 100 includes an input unit (for example, a keyboard and a mouse) that receives various operations from an administrator who uses the learning device 100, and a display unit (for example, a liquid crystal display) that displays various types of information. You may have.

(About the communication unit 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 in a wired or wireless manner, and transmits / receives information to / from the user terminal 10, the advertiser terminal 20, and the web server 30 via the network N.

(About the storage unit 120)
The storage unit 120 is realized by, for example, a semiconductor memory device such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disk. The storage unit 120 includes an advertisement information storage unit 121, a terminal information storage unit 122, and a learning model storage unit 126.

(Regarding the advertisement information storage unit 121)
The advertisement information storage unit 121 stores information related to advertisements submitted from the advertiser terminal 20. Here, FIG. 6 illustrates an example of the advertisement information storage unit 121 according to the embodiment. FIG. 6 is a diagram illustrating an example of the advertisement information storage unit 121 according to the embodiment. In the example illustrated in FIG. 6, the advertisement information storage unit 121 includes items such as “advertiser ID”, “advertisement ID”, and “corresponding application ID”.

  The “advertiser ID” indicates identification information for identifying the advertiser or the advertiser terminal 20. “Advertisement ID” indicates identification information for identifying an advertisement submitted by an advertiser. The “corresponding application ID” indicates identification information for identifying the application associated with the advertisement.

  In the present specification, identification information as shown in FIG. 6 may be used as a reference symbol. For example, the advertiser identified by the advertiser ID “B10” is identified by “advertiser B10”, the advertisement identified by the advertisement ID “C10” is identified by “advertisement C10”, and (corresponding) the application ID “A10”. May be referred to as “application A10”.

  That is, the example of the data illustrated in FIG. 6 indicates that the advertiser B10 identified by the advertiser ID “B10” has submitted the advertisement identified by the advertisement ID “C10” or “C11”. . In addition, the application associated with the advertisement C10 indicates that the application A10 is identified by the corresponding application ID “A10”.

  Note that an application may be associated with a plurality of advertisements instead of being associated with an advertisement on a one-to-one basis. For example, as shown in FIG. 6, the application A20 is associated with an advertisement C20 and an advertisement C21. This indicates that the advertisement C20 and the advertisement C21 are targeted for the same application A20, although the contents of the advertisement image data and the like are different.

  Note that the content data (text data, moving image content, and still image content) of the advertisement that is actually distributed to the user terminal 10 may be stored in a predetermined storage server provided separately from the learning device 100. In this case, the learning device 100 identifies an advertisement stored in an external storage server based on the advertisement ID stored in the advertisement information storage unit 121. Then, the learning device 100 controls the storage server to distribute the specified advertisement to the user terminal 10.

  The advertisement information storage unit 121 may store other information related to advertisements. For example, the advertisement information storage unit 121 may store distribution destination conditions specified for each advertisement, the number of distributions specified for each advertisement (specified number of impressions), and the like. In addition, the advertisement information storage unit 121 may store an index value indicating an advertisement effect. For example, the advertisement information storage unit 121 may store an index value such as CPI (Cost Per Install) or CTR (Click Through Rate) for each advertisement.

(Terminal information storage unit 122)
The terminal information storage unit 122 stores information related to the user terminal 10 that is an advertisement distribution target. As shown in FIG. 5, the terminal information storage unit 122 includes an attribute table 123, a device table 124, and an application table 125 as data tables for storing terminal information.

(About attribute table 123)
Here, FIG. 7 shows an example of the attribute table 123 according to the embodiment. FIG. 7 is a diagram illustrating an example of the attribute table 123 according to the embodiment. The attribute table 123 mainly stores information related to attributes of users who use the user terminal 10. In the example illustrated in FIG. 7, the attribute table 123 includes items such as “terminal ID”, “sex”, “age”, and “region”.

  “Terminal ID” is identification information for identifying the user terminal 10. “Gender” indicates the gender of the user who uses the user terminal 10. “Age” indicates the age of the user who uses the user terminal 10. “Region” indicates the residence of the user who uses the user terminal 10. The “region” may store a region name (such as Kanto region) indicating a certain range corresponding to the user's residence, a country name, or the like, instead of a specific address.

  That is, in the example of the data illustrated in FIG. 7, the gender of the user of the user terminal 10 identified by the terminal ID “F11” is “male”, the age is “30s”, and the area where the user resides is “ Indicates "A prefecture".

  The attribute information stored in the attribute table 123 is not necessarily accurate information. For example, the learning device 100 may store “estimated gender”, “estimated age”, and the like estimated from the device information in the attribute table 123. In addition to the “region”, the attribute table 123 may store information such as “city rank” indicating the city degree of the region.

(About the device table 124)
Next, FIG. 8 shows an example of the device table 124 according to the embodiment. FIG. 8 is a diagram illustrating an example of the device table 124 according to the embodiment. The device table 124 mainly stores device information indicating information of the device itself called the user terminal 10. In the example illustrated in FIG. 8, the device table 124 includes items such as “terminal ID”, “model number”, “brand name”, “days since release”, “communication carrier”, “manufacturer name”, and “resolution”. Have.

  The “terminal ID” corresponds to the same item shown in FIG. “Model number” indicates the model number of the user terminal 10. “Brand name” indicates a brand name assigned to the user terminal 10. “Elapsed days after release” indicates the number of days that have elapsed since the user terminal 10 was released. “Communication carrier” indicates the company name of the communication carrier that provides the communication line of the user terminal 10. “Resolution” indicates the screen resolution of the user terminal 10.

  That is, in the example of the data illustrated in FIG. 8, the user terminal 10 identified by the terminal ID “F11” indicates that the model number is “XX-YY01” and the brand name is “AAA”. In addition, the user terminal 10 identified by the terminal ID “F11” has “336 days” after the release, the communication carrier is “BBB company”, the manufactured manufacturer is “CCC company”, and the resolution is Indicates “1280 × 720”.

(About the application table 125)
Next, FIG. 9 shows an example of the application table 125 according to the embodiment. FIG. 9 is a diagram illustrating an example of the application table 125 according to the embodiment. The application table 125 mainly stores information related to applications installed in the user terminal 10. In the example illustrated in FIG. 9, the application table 125 includes “terminal ID”, “number of installed applications”, “number of non-game apps”, “number of game apps”, “number of new apps”, “number of old apps”, “ It has an item such as “installed application ID”.

  The “terminal ID” corresponds to the same item shown in FIG. “Number of installed applications” indicates the total number of applications installed in the user terminal 10. The “number of non-game apps” indicates the number of apps other than game apps among installed apps. “Number of game applications” indicates the number of game applications among the installed applications. The “number of new applications” indicates the number of applications that have been provided relatively recently (for example, within one year) among the installed applications. “The number of old apps” indicates the number of apps other than the new apps among the installed apps. “Installed application ID” indicates identification information of each application installed in the user terminal 10.

  That is, the example of the data illustrated in FIG. 9 indicates that “35” applications are installed in the user terminal 10 identified by the terminal ID “F11”. In addition, among the applications installed on the user terminal 10 identified by the terminal ID “F11”, the number of non-game applications is “22”, and the number of game applications is “13”. Further, among the applications installed on the user terminal 10 identified by the terminal ID “F11”, the number of new applications is “15” and the number of old applications is “20”. The application installed on the user terminal 10 identified by the terminal ID “F11” is identified by identification information such as “A101”, “A103”, “A107”, “A108”, “A122”, and the like. Indicates that this is an app.

(About the learning model storage unit 126)
The learning model storage unit 126 stores information regarding learning processing of the learning device 100. As illustrated in FIG. 5, the learning model storage unit 126 includes a setting table 127, a variable table 128, and a model table 129 as data tables for storing information related to learning processing.

(About setting table 127)
The setting table 127 mainly stores information related to the setting of terminal information and variables extracted from the terminal information. The setting table 127 further includes a plurality of tables for each type of stored information. The setting content stored in the setting table 127 is automatically set based on, for example, an input from the administrator of the learning device 100 or information extracted from a site or the like that introduces the application, as will be described later. It is done.

  Here, FIG. 10 shows an example of the setting table 127 according to the embodiment. FIG. 10 is a diagram (1) illustrating an example of the setting table 127 according to the embodiment. In FIG. 10, as an example of the setting table 127, a variable setting table 127 </ b> A in which associations between variable identification information used for learning and variable types and contents are set. In the example shown in FIG. 10, the variable setting table 127A has items such as “variable ID”, “type”, and “variable name”.

  “Variable ID” indicates identification information for identifying a variable. “Type” indicates a type corresponding to the variable ID. The type is used to classify variables. “Variable name” indicates the name of each variable subdivided into types.

  That is, in the example of the data illustrated in FIG. 10, the variable identified by the variable ID “E26” is a variable related to the type “interest of interest”, and in particular, is a variable indicating interest in “internet shopping”. Show. The variable identified by the variable ID “E58” is a variable related to the type “game preference”, and particularly indicates a variable indicating interest in “shooting”.

  Next, FIG. 11 shows another example of the setting table 127 according to the embodiment. FIG. 11 is a diagram (2) illustrating an example of the setting table 127 according to the embodiment. FIG. 11 illustrates an application setting table 127B in which variables associated with applications are set as an example of the setting table 127. In the example illustrated in FIG. 11, the application setting table 127 </ b> B includes items such as “application ID”, “type”, “element”, and “variable name”.

  “Application ID” indicates identification information for identifying an application. “Type” indicates the type of application. “Element” indicates an element included in the application. “Variable name” indicates a variable set based on the type or element of the application.

  That is, in the example of the data illustrated in FIG. 11, the application A51 identified by the application ID “A51” is an application belonging to the type “communication system”, and “communication”, “contact”, “chat” Or an application having an element indicated by a keyword such as “SNS”. For this reason, it is shown that variables identified by names such as “interest_lifestyle”, “interest_SNS”, and “interest_encounter” are set in the application A51.

(Variable table 128)
Next, FIG. 12 shows an example of the variable table 128 according to the embodiment. FIG. 12 is a diagram illustrating an example of the variable table 128 according to the embodiment. As illustrated in FIG. 12, the variable table 128 includes items such as “processing target application ID”, “terminal ID”, “target application installation (purpose variable)”, and “feature information data (explanatory variable)”.

  “Processing target application ID” indicates identification information for identifying an application for which a model is generated. “Terminal ID” indicates identification information for identifying the user terminal 10.

  “Target application installation (objective variable)” is information that becomes an objective variable in a model generated by the learning apparatus 100. Specifically, “target application installation (purpose variable)” is information indicating whether or not an application to be processed has been installed in the user terminal 10. In the example of FIG. 12, if the item “target application installation (object variable)” is “1”, it indicates that the processing target application is installed in the user terminal 10. On the other hand, if the item “target application installation (objective variable)” is “0”, it indicates that the processing target application is not installed in the user terminal 10.

  “Characteristic information data (explanatory variable)” is information that becomes an explanatory variable in a model generated by the learning apparatus 100. In the example of FIG. 12, “G11” is conceptually described, but actually, “feature information data (explanatory variable)” is data in which a plurality of explanatory variables of the user terminal 10 are listed. It is. That is, the feature information data is a collection of feature information extracted from the terminal information of the user terminal 10, and is data indicating the features of the user terminal 10 (in other words, a user who uses the user terminal 10).

  That is, in the example of the data illustrated in FIG. 12, the application A100 identified by the processing target application ID “A100” is identified by the user terminal 10 identified by the terminal ID “F11” and the terminal ID “F13”. It shows that it was installed in the user terminal 10. On the other hand, the application A100 is not installed in the user terminal 10 identified by the terminal ID “F12”. The feature information data of the user terminal 10 identified by the terminal ID “F11” is “G11”, the feature information data of the user terminal 10 identified by the terminal ID “F12” is “G12”, and the terminal ID The feature information data of the user terminal 10 identified by “F13” indicates “G13”.

(About model table 129)
Next, FIG. 13 shows an example of the model table 129 according to the embodiment. FIG. 13 is a diagram illustrating an example of the model table 129 according to the embodiment. As illustrated in FIG. 13, the model table 129 includes items such as “model ID” and “corresponding application ID”.

  “Model ID” indicates identification information for identifying a model. “Corresponding application ID” indicates identification information for identifying an application corresponding to the model.

  That is, in the example of the data illustrated in FIG. 13, the model M100 identified by the model ID “M100” is a model corresponding to the application ID “A100”.

(About the control unit 130)
The control unit 130 is configured such that, for example, a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like stores various programs (corresponding to an example of a learning program) stored in a storage device inside the learning device 100 as a work area. This is realized by being executed as The control unit 130 is realized by an integrated circuit such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA).

  As illustrated in FIG. 5, the control unit 130 includes a submission receiving unit 131, an acquisition unit 132, a learning unit 133, an extraction unit 134, a generation unit 135, a reception unit 136, a specification unit 137, and distribution. And implements or executes the functions and operations of information processing described below. Note that the internal configuration of the control unit 130 is not limited to the configuration illustrated in FIG. 5, and may be another configuration as long as information processing described later is performed. In addition, the connection relationship between the processing units included in the control unit 130 is not limited to the connection relationship illustrated in FIG. 5, and may be another connection relationship.

(About the submission acceptance part 131)
The submission accepting unit 131 accepts an advertisement submission from the advertiser terminal 20. Then, the submission accepting unit 131 associates an advertiser ID for identifying the advertiser of the submission source, an advertisement ID for identifying the advertisement, and a corresponding application ID for identifying the application corresponding to the advertisement, thereby providing advertisement information. Store in the storage unit 121.

  The submission accepting unit 131 may accept from the advertiser the number of distributions specified for each advertisement, conditions relating to the user of the distribution destination, and the like. In addition, for example, when the advertisement is distributed without using a model corresponding to the advertisement (that is, the application), the submission accepting unit 131 includes a targeting element such as designation of an attribute of a user to which the advertisement is distributed. May be accepted.

(About the acquisition unit 132)
The acquisition unit 132 acquires various types of information. For example, the acquisition unit 132 acquires terminal information that is information regarding the user terminal 10 that receives the advertisement. Specifically, the acquisition unit 132 acquires attribute information of a user who uses the user terminal 10 as terminal information. Moreover, the acquisition part 132 acquires the apparatus information which is the information of the apparatus itself of the user terminal 10 as terminal information. Further, the acquisition unit 132 acquires application information that is information related to an application installed in the user terminal 10 as terminal information.

  Moreover, the acquisition part 132 may acquire the information regarding the received advertisement as the terminal information, and the information regarding the user's behavior with respect to the advertisement. For example, when the user selects (for example, clicks) an advertisement, the acquisition unit 132 acquires information regarding the web page on which the advertisement is displayed and the location of the advertisement space. The acquisition unit 132 may acquire time information such as how much time has passed since the user finally selected an arbitrary advertisement. In addition, the acquisition unit 132 may acquire information such as a contact time period for an advertisement such as which time period of the day the user often selects an advertisement. Thus, the information acquired by the acquisition unit 132 may be used as an explanatory variable of the user terminal 10 by the learning unit 133 described later.

  Moreover, the acquisition part 132 may acquire the result information which shows whether the user terminal 10 which received the advertisement performed the predetermined | prescribed action regarding an advertisement. Specifically, the acquisition unit 132 acquires result information indicating whether or not the user terminal 10 has installed an application to be subjected to learning processing by the learning unit 133 described later.

  The acquisition unit 132 may acquire terminal information and result information at any timing. For example, the acquisition unit 132 may acquire terminal information or the like when an advertisement distribution request is received from the user terminal 10. In addition, when the application associated with the advertisement is installed in the user terminal 10, if the program is such that a predetermined communication is sent from the user terminal 10 side to the learning device 100 side, the acquisition unit 132 is When receiving such communication, terminal information and the like are acquired. Alternatively, the acquisition unit 132 may acquire terminal information of the user terminal 10 from an external server at an arbitrary timing, instead of acquiring terminal information from the user terminal 10.

  Then, the acquisition unit 132 stores the acquired information in a predetermined storage unit. For example, when acquiring the terminal information, the acquiring unit 132 stores the acquired information in the terminal information storage unit 122. Alternatively, the acquisition unit 132 may send the acquired information to a processing unit such as the learning unit 133.

  Note that the acquisition unit 132 may be realized by various known methods for acquiring information related to the advertisement, such as whether the advertisement is clicked or whether an application corresponding to the advertisement is installed. For example, the acquisition unit 132 may acquire information related to the advertisement using a notification function realized by a web beacon or the like.

(About the learning unit 133)
The learning unit 133 learns the relationship between the terminal information and the result information indicating whether or not the predetermined action related to the predetermined information content has been executed. And the learning part 133 produces | generates the model which shows the tendency which performs the predetermined | prescribed action regarding a predetermined | prescribed information content based on the learning result. Further, the learning unit 133 acquires the terminal information even after generating the model, and optimizes the model by continuing the learning of the model. The learning unit 133 includes an extraction unit 134 and a generation unit 135, and realizes the above processing.

(About the extraction unit 134)
The extraction unit 134 extracts feature information that characterizes the user terminal 10 from the terminal information acquired by the acquisition unit 132. For example, the extraction unit 134 extracts information such as the brand name of the user terminal 10 and the number of days elapsed after sale based on the device information.

  Further, the extraction unit 134 extracts feature information that characterizes the user terminal 10 (that is, the user of the user terminal 10) based on the application information. For example, the extraction unit 134 specifies an application installed in the user terminal 10. Then, the extraction unit 134 identifies the type and element of the application set for the application. Further, the extraction unit 134 identifies a variable that is set based on the type or element of the application. That is, the extraction unit 134 specifies a variable used in a model to be described later as feature information extracted from the terminal information.

  For example, the extraction unit 134 extracts feature information from the terminal information based on settings input by the administrator of the learning device 100. Taking FIG. 11 as an example, the extraction unit 134 specifies that variables such as “interest_lifestyle”, “interest_SNS”, and “interest_encounter” are associated with the application A51. Then, when the application A51 is specified as an application installed in the user terminal 10 to be processed, the extraction unit 134 extracts the above variables as information characterizing the user terminal 10.

  Note that the extraction unit 134 may extract feature information using information other than the setting input by the administrator of the learning device 100. For example, in an application download site or the like, there is a case where text data that is an introductory sentence about an application is associated with each application. In this case, the extraction unit 134 analyzes the text data to extract a keyword that is an application type or an element. For example, it is assumed that the extracting unit 134 extracts nouns such as “communication” and “contact” by performing morphological analysis in the text data associated with the application A51. In this case, the extraction unit 134 sets the extracted noun as a keyword as an element of the application A51. Furthermore, the extraction unit 134 sets a variable associated with the word set as an element as a variable corresponding to the application A51. Thereby, the extraction unit 134 can associate a variable (feature information) with an application without using an artificial process such as an administrator of the learning apparatus 100.

(About the generator 135)
The generation unit 135 learns the relationship between the result information indicating whether or not the predetermined action related to the predetermined information content has been executed and the terminal information acquired by the acquisition unit 132, thereby determining the predetermined information related to the predetermined information content. Generate a model that shows the tendency of the action to be executed. Specifically, the generation unit 135 generates a model indicating a tendency to install a predetermined application by learning the relationship between the result information indicating whether or not the predetermined application has been installed and the terminal information. .

  As described above, the generation unit 135 generates a model for each application (in the case where the application and the advertisement correspond one-to-one, it can also be referred to as “for each advertisement”). That is, the generation unit 135 uses result information indicating whether or not a predetermined application is installed as an objective variable in machine learning. And the production | generation part 135 makes the feature information in the user terminal 10 which received the advertisement matched with the predetermined | prescribed application the explanatory variable in machine learning. And the production | generation part 135 produces | generates the model regarding a predetermined application using an objective variable and an explanatory variable. Thereby, the production | generation part 135 produces | generates the model which can pinpoint the user terminal 10 which tends to install a predetermined application with high precision.

  The generation unit 135 generates a model using various explanatory variables. For example, the generation unit 135 uses user feature information extracted from an application installed in the user terminal 10 as an explanatory variable. As an example, the generation unit 135 is information specified based on the type or element set in the application installed in the user terminal 10 as the feature information extracted from the application information, and the interest of the user of the user terminal 10 A model is generated using interest information, which is information indicating. Further, the generation unit 135 is information specified based on the type or element set in the game application installed in the user terminal 10 as feature information extracted from the application information, and the game of the user of the user terminal 10 A model may be generated using preference information which is information indicating preference. As described above, the type or element of the application may be set for each application by the extraction unit 134 analyzing the introduction sentence such as the download site. That is, the generation unit 135 can set the type or element of the application installed in the user terminal 10 based on the introduction of the application set for the application from the provider side that provides the application.

  Hereinafter, an example of a model generated by the generation unit 135 will be described. Note that the learning method related to the model generated by the generation unit 135 is not limited to the following example, and various known machine learning methods may be employed.

  For example, the generation unit 135 generates an expression indicating a relationship between whether or not a predetermined application is installed in the user terminal 10 and the terminal information of the user terminal 10. Furthermore, the generation unit 135 calculates what weight each piece of feature information extracted from the terminal information of the user terminal 10 has with respect to an event that a predetermined application is installed. Accordingly, the generation unit 135 can obtain information such as how much individual feature information contributes to an event of installation of a predetermined application. For example, the generation unit 135 generates the following equation (1) when generating a model indicating a tendency to install the application A100 illustrated in FIG.

y _(A100) = ω ₁ · x ₁ + ω ₂ · x ₂ + ω ₃ · x ₃ ... + ω _N · x _N (1) (N is an arbitrary number)

The above formula (1) is created for each user terminal 10 to which an advertisement associated with the application A100 is distributed, for example. In the above formula (1), “y _(A100) ” indicates an event “whether or not the application A100 is installed in the user terminal 10 to which the advertisement is distributed”.

In the above equation (1), “x” corresponds to each explanatory variable of the user terminal 10. Further, the explanatory variables “x ₁ , x ₂ , x ₃ ,..., X _N ” in the above formula (1) correspond to the variables shown in FIG. Specifically, if it contains "x _26" in the above formula (1), "x _26" corresponds to the variable of the variable ID "E26" shown in FIG. 10, the contents of "interest _ Net Shopping ". That is, the right side of the above formula (1) corresponds to feature information data extracted from the terminal information of the user terminal 10 as shown in FIG.

In the above formula (1), “ω” is a coefficient of “x” and indicates a predetermined weight value. Specifically, “ω ₁ ” is a weight value of “x ₁ ”, “ω ₂ ” is a weight value of “x ₂ ”, and “ω ₃ ” is a weight value of “x ₃ ”. It is. As described above, the above equation (1) is a variable including the explanatory variable “x” corresponding to the feature information extracted from the terminal information and the predetermined weight value “ω” (for example, “ω ₁ · x _1”). ).

A specific example will be described with reference to FIGS. In the following description, the terminal ID is used as a user's reference code for convenience. For example, the user of the user terminal 10 identified by the terminal ID “F11” is represented as “user F11”. In the above formula (1), “x ₁ ” corresponds to “gender”, “x ₂ ” corresponds to “age”, and “x ₃ ” is an explanatory variable corresponding to “region”. And In this case, the above formula (1) corresponding to the user F11 can be expressed as the following formula (2).

y _{(A100, F11)} (= 1) = ω ₁ · (male) + ω ₂ · (30's) + ω ₃ · (A prefecture) + ω _N · x _N (2)

  Similarly, the user F12 and the user F13 can be expressed as the following formulas (3) and (4), respectively.

y _{(A100, F12)} (= 0) = ω ₁ · (male) + ω ₂ · (40 years old) + ω ₃ · (B prefecture) + ω _N · x _N (3)
y _{(A100, F13)} (= 1) = ω ₁ · (female) + ω ₂ · (20s) + ω ₃ · (C prefecture) + ω _N · x _N (4)

The generation unit 135 generates an expression for each user (for each user terminal 10) as in the expressions (2), (3), and (4), and uses the generated expression as a machine learning sample. Then, the generation unit 135 derives a value corresponding to the predetermined weight value “ω” by performing calculation processing of the sample expression. That is, the generation unit 135 determines a predetermined weight value “ω” that satisfies the expressions (2), (3), and (4). In other words, the generation unit 135 can determine the weight value “ω” indicating the influence of the predetermined explanatory variable on the objective variable “y”. For example, as information indicating the characteristics of the user who installs the application A100, if “gender” greatly contributes compared to other variables, the value of the weight value “ω ₁ ” corresponding to “sex” is A large value is calculated in comparison with other variables. In the above example, the explanatory variable is an example of user attribute information such as “gender”, “age”, and “region”. Various explanatory variables such as 10 device information and application information are included. That is, the generation unit 135 can generate a model using terminal information or various feature information extracted from the terminal information as illustrated in FIGS. 7 to 12 as explanatory variables.

  As described above, the generation unit 135 generates a model that associates the event that the application A 100 is likely to be installed with the terminal information acquired from the user terminal 10. In the calculation process using the above formulas (2), (3), and (4), the left side is not set to “1” or “0” itself, but a predetermined error is assumed and the difference from the error is assumed. The optimal solution of “ω” may be calculated using a method such as a least square method that approximates the squared value to the minimum value.

  In addition, when substituting the feature information extracted from the terminal information of the user terminal 10 for the generated model, the generation unit 135 determines whether it is “present” or “none” such as “interest_net shopping”. For the variable, a numerical value “1” or “0” is substituted. Further, the generation unit 135 may apply various known methods such as normalization so that an event indicated as an explanatory variable can be handled by a model according to a known method with respect to variables such as resolution. .

  In addition, after generating the model, the generation unit 135 may update the model as needed when the acquisition unit 132 acquires new terminal information. Thereby, the production | generation part 135 can optimize the model which showed the characteristic of the user terminal 10 which tends to install an application.

(Receiving unit 136)
The receiving unit 136 receives an advertisement distribution request. Specifically, the receiving unit 136 receives a request related to distribution of an advertisement displayed in an advertisement space included in the web page, which is a request transmitted from the user terminal 10 that displays the web page.

  The receiving unit 136 may receive a request for providing information transmitted from the user terminal 10 and receive terminal information from the user terminal 10. For example, the receiving unit 136 receives identification information for identifying the user terminal 10 as an example of the terminal information of the user terminal 10. In this case, the reception unit 136 sends the received information to the acquisition unit 132 and the specification unit 137.

(About the specific part 137)
Based on the model generated by the generation unit 135, the specification unit 137 specifies the user terminal 10 that is a providing destination that provides predetermined information content. Specifically, the specifying unit 137 displays the advertisement associated with the application based on the tendency (that is, the output value) indicated by the model indicating the tendency that the application associated with the advertisement is installed. The user terminal 10 that is a delivery destination to be delivered is specified.

  For example, when the receiving unit 136 receives an advertisement distribution request, the specifying unit 137 acquires terminal information of the user terminal 10 that has transmitted the request. Then, the specifying unit 137 inputs the acquired terminal information into the model generated by the generating unit 135. And the specific | specification part 137 determines whether the advertisement corresponding to a model is delivered based on the result output from a model. For example, the specifying unit 137 calculates the probability that the user installs the application based on the result output from the model. Then, when the calculated probability indicates a predetermined value or more, the specifying unit 137 determines that the advertisement corresponding to the model is distributed, and specifies the user terminal 10 that is the distribution destination.

  As another example, when the value output from the model exceeds a predetermined threshold, the specifying unit 137 determines that there is a high possibility that the application corresponding to the model is installed in the user terminal 10 that is the distribution destination. To do. And the specific | specification part 137 specifies the delivery destination which delivers the advertisement corresponding to the said application. On the other hand, when the value output from the model does not exceed the predetermined threshold, the specifying unit 137 determines that there is a low possibility that the application corresponding to the model is installed in the user terminal 10 that is the distribution destination. In this case, for example, the specifying unit 137 may repeatedly input the terminal information to another model until the value output from the model exceeds a predetermined threshold. Alternatively, the specifying unit 137 may determine that the advertisement is distributed without using the model. The specifying unit 137 may determine that the advertisement is distributed without using the model even when there is no model corresponding to the advertisement to be distributed. In the case of the above example, the value output by the model is not limited to the probability of installing the application, and may be any index value.

  Further, even after the model is generated, the specifying unit 137 may specify the user terminal 10 that distributes the advertisement without using the model in a predetermined case. For example, when the learning process is excessively performed, the specifying unit 137 may determine to always distribute the same advertisement to a certain user terminal 10. In such a case, the specifying unit 137 may perform the specifying process without using the model in order to ensure a certain randomness in the advertisement distribution.

(About distribution unit 138)
The distribution unit 138 distributes the advertisement to the user terminal 10 determined as the advertisement distribution destination based on the result determined by the specifying unit 137.

  As described above, the advertisement data itself that is actually distributed to the user terminal 10 may not be stored in the advertisement information storage unit 121 of the learning device 100. For example, the distribution unit 138 may distribute the advertisement to the user terminal 10 by transmitting a control command to a predetermined storage server provided outside.

[4. Configuration of terminal device]
Next, the configuration of the user terminal 10 according to the embodiment will be described with reference to FIG. FIG. 14 is a diagram illustrating a configuration example of the user terminal 10 according to the embodiment. As illustrated in FIG. 14, the user terminal 10 includes a communication unit 11, an input unit 12, a display unit 13, a detection unit 14, a storage unit 15, and a control unit 16. Note that the connection relationship between the processing units included in the user terminal 10 is not limited to the connection relationship illustrated in FIG. 14, and may be another connection relationship.

  The communication unit 11 is connected to the network N by wire or wireless, and transmits and receives information to and from the web server 30 and the learning device 100. For example, the communication unit 11 is realized by a NIC or the like.

  The input unit 12 is an input device that receives various operations from the user. For example, the input unit 12 is realized by an operation key or the like provided in the user terminal 10. The input unit 12 may include an imaging device (such as a camera) for capturing an image and a sound collection device (such as a microphone) that collects sound.

  The display unit 13 is a display device for displaying various information. For example, the display unit 13 is realized by a liquid crystal display or the like. In addition, when a touch panel is employ | adopted for the user terminal 10, a part of input part 12 and the display part 13 are integrated.

  The detection unit 14 detects various operations on the user terminal 10 and environmental information around the user terminal 10. For example, the detection unit 14 is realized by a sensor or an antenna that detects various types of information. Specifically, the detection unit 14 includes a communication status related to a device connected to the user terminal 10, illuminance and noise around the user terminal 10, physical movement of the user terminal 10, position information of the user terminal 10, and the like. Is detected.

  The storage unit 15 stores various information. The storage unit 15 is realized by, for example, a semiconductor memory device such as a RAM or a flash memory, or a storage device such as a hard disk or an optical disk. In the example illustrated in FIG. 14, the storage unit 15 includes an installation information storage unit 151. The installation information storage unit 151 stores, for example, information on applications installed in the user terminal 10.

  The control unit 16 is realized, for example, by executing various programs stored in the storage device inside the user terminal 10 using the RAM as a work area by a CPU, an MPU, or the like. The control unit 16 is realized by an integrated circuit such as ASIC or FPGA, for example.

  The control unit 16 controls various processes performed in the user terminal 10. As illustrated in FIG. 14, the control unit 16 includes a reception unit 161, an acquisition unit 162, an execution unit 163, and a transmission unit 164, and realizes or executes functions and operations of information processing described below. .

  The receiving unit 161 receives various information. For example, the receiving unit 161 receives information transmitted from the web server 30 or the learning device 100. Specifically, the receiving unit 161 receives an advertisement distributed in response to an advertisement distribution request. The receiving unit 161 receives various information detected by the detecting unit 14.

  The acquisition unit 162 acquires various information and data. For example, the acquisition unit 162 acquires a web page that the user desires to browse by accessing the web server 30. The acquisition unit 162 acquires the advertisement data received by the reception unit 161. In addition, the acquisition unit 162 acquires data used for application installation via an application download site or the like.

  The execution unit 163 executes various processes in the user terminal 10. For example, the execution unit 163 executes processing for installing an application. When the execution unit 163 installs an application, information related to installation is stored in the installation information storage unit 151.

  The transmission unit 164 transmits various information. For example, when the advertising space is included in the web page acquired by the acquiring unit 162, the transmitting unit 164 transmits an advertisement distribution request to the learning device 100. The transmission unit 164 transmits the terminal information of the user terminal 10 to the learning device 100 with reference to the storage unit 15 and the like.

[5. Processing procedure)
Next, a processing procedure performed by the learning apparatus 100 according to the embodiment will be described with reference to FIGS. 15 and 16. First, a processing procedure related to model generation will be described with reference to FIG. FIG. 15 is a flowchart (1) illustrating a processing procedure according to the embodiment.

  As illustrated in FIG. 15, the learning device 100 receives an advertisement submission from the advertiser terminal 20 (step S <b> 101). Then, the learning device 100 specifies an application corresponding to the submitted advertisement (step S102).

  Thereafter, the learning device 100 determines whether an advertisement distribution request has been received from the user terminal 10 (step S103). If the learning device 100 has not received a request for advertisement distribution, the learning device 100 waits until the request is received (step S103; No).

  On the other hand, when the request for advertisement distribution is received (step S103; Yes), the learning device 100 distributes the advertisement to the user terminal 10 that has transmitted the request (step S104). Thereafter, the learning device 100 acquires result information indicating whether or not an application corresponding to the advertisement has been installed and terminal information from the user terminal 10 that has distributed the advertisement (step S105).

  And the learning apparatus 100 extracts feature information from terminal information (step S106). Subsequently, the learning apparatus 100 determines whether a model corresponding to the application to be processed already exists (step S107). If the model exists (step S107; Yes), the learning device 100 performs a model update process based on the acquired terminal information (step S108).

  On the other hand, when the model does not exist (step S107; No), the learning device 100 newly performs a model generation process based on the acquired result information and terminal information (step S109). Thereafter, the learning device 100 optimizes the model by repeating the terminal information acquisition process. Note that the learning device 100 may omit the processes of steps S101 to S104 when the result information and the terminal information can be acquired from an external server or the like.

  Next, a processing procedure related to advertisement distribution will be described with reference to FIG. FIG. 16 is a flowchart (2) illustrating a processing procedure according to the embodiment.

  As illustrated in FIG. 16, the learning device 100 determines whether an advertisement distribution request has been received from the user terminal 10 (step S201). If the learning device 100 has not received a request for advertisement distribution, the learning device 100 waits until the request is received (step S201; No).

  On the other hand, when a request for advertisement distribution is received (step S201; Yes), the learning device 100 acquires terminal information from the user terminal 10 that transmitted the request (step S202). Then, the learning device 100 inputs the acquired terminal information as an explanatory variable to an arbitrarily selected model (step S203).

  Then, the learning device 100 determines whether or not the output value has exceeded a certain threshold value (step S204). When the predetermined threshold value is exceeded (step S204; Yes), the learning apparatus 100 delivers an advertisement corresponding to the model to the user terminal 10 (step S205).

  On the other hand, when the output value does not exceed a certain threshold (step S204; No), the learning device 100 determines whether another model exists (step S206). When there is no other model (step S206; No), the learning device 100 distributes the advertisement without using the model (step S207).

  On the other hand, when another model exists (step S206; Yes), the learning apparatus 100 selects another model (step S208). And the learning apparatus 100 repeats the process which inputs terminal information into the selected model (step S203).

[6. (Modification)
The learning device 100 described above may be implemented in various different forms other than the above embodiment. Therefore, in the following, another embodiment of the learning device 100 will be described.

[6-1. (Variable type)
In embodiment mentioned above, the acquisition part 132 showed the example which acquires the attribute information of the user of the user terminal 10, the apparatus information of the user terminal 10, and application information as terminal information. In addition, the extraction unit 134 has shown an example in which feature information is extracted based on the information acquired by the acquisition unit 132. The acquisition unit 132 and the extraction unit 134 are not limited to these examples, and may acquire various terminal information or extract feature information.

  For example, the acquisition unit 132 may acquire the OS (Operating System) type and version information of the user terminal 10, the resolution of the vertical screen and the horizontal screen, the total number of pixels, and the like. In addition, the acquisition unit 132 may acquire the proportion of the game application among the applications installed in the user terminal 10.

  Further, the extracting unit 134 may extract the feature information extracted from the terminal information in addition to the variables exemplified in FIG. For example, if the type of feature information to be extracted is an interest, the extraction unit 134 may be a part-time job, a job change, an encounter or a marriage, a house or a car, a household account book or a coupon or a point, a variety of sports such as golf, Various information such as English language and cooking, travel, fashion, brand, diet, female target products, student target products, flea market, illustrations and photographs may be extracted as feature information (explanatory variables). For example, the extraction unit 134 extracts variables corresponding to various interests based on elements set for each application.

  Further, the extraction unit 134 may be a genre other than the genre illustrated in FIG. 10 (for example, sports, casual, board, romance, beautiful girl, breeding, etc.) if it is game preference as the type of feature information extracted from the terminal information. (Sengoku / Sangokushi), classification such as card, strategy, scenario, gambling, character, presence / absence of charge, etc., may be extracted as feature information (explanatory variable). That is, according to the extraction unit 134, information indicating that the user of the user terminal 10 has installed a game application that is given a variable such as “game preference_card” or “game preference_strategy”, for example. Can be extracted. And the specific | specification part 137 outputs the information of whether the application matched with the model tends to be installed from the said user by inputting the extracted information into a model as an explanatory variable. it can.

[6-2. (Type of information content)
In the embodiment described above, an advertisement associated with an application is taken as an example of information content provided to a user. However, information content is not limited to such advertisements. For example, the information content may be an advertisement that is not associated with an application and that displays a landing page when selected. In this case, the generation unit 135 may generate a model in which the objective variable is “an advertisement is clicked”. Thereby, the specific | specification part 137 can identify the user who tends to click the advertisement distributed and can distribute an advertisement.

  The information content is not limited to advertisements, and may be product recommendation information. In this case, the generation unit 135 may generate a model in which the objective variable is “to purchase recommended products”. Thereby, the specific | specification part 137 can identify the user who tends to purchase the recommended goods, and can provide recommendation information.

[6-3. (Relation with media)
In the embodiment described above, the specifying unit 137 has shown an example in which the user terminal 10 that distributes the advertisement is specified based on the model generated by the generating unit 135. Here, the specifying unit 137 may further perform a specifying process based on a predetermined condition received from the advertiser.

  For example, some advertisers may specify medium content (for example, a web page or an application) on which an advertisement submitted by the advertiser is displayed. Specifically, there is a case where an advertiser wants to display his / her advertisement on content on which information of a specific category is posted in order to increase the appeal effect of the advertisement. Alternatively, the advertiser may want to prevent an advertisement that he / she submits from being displayed on a web page provided by a competitor.

  In this case, the specifying unit 137 may specify the user terminal 10 that distributes the advertisement in consideration of the result output by the model and the condition specified by the advertiser. As a result, the advertiser may not display an advertisement on a web page that is not desired to have its own advertisement even if the web page is displayed by a user who is likely to be installed. it can.

[6-4. Use of external information)
When the learning apparatus 100 can generate a model by using more detailed information regarding the user who uses the user terminal 10 as the terminal information, the learning apparatus 100 may use the information. For example, the learning apparatus 100 acquires a public survey result or the like published by a public institution or the like regarding the user. As an example, the learning apparatus 100 acquires a survey result such as a relationship between an average annual income of a general citizen including a user and a region. Then, when the user's area can be specified, the learning device 100 may use the average annual income in that area as one of the feature information.

  As another example, the learning apparatus 100 acquires a survey result such as a relationship between an average time spent for public entertainment including a user and childcare and age. If the age of the user can be specified or estimated, the learning device 100 may use an average time spent for entertainment or childcare corresponding to the age as one of the feature information. Through such processing, the learning apparatus 100 can generate a model using information that is difficult to acquire between terminals. Thereby, the learning apparatus 100 can obtain feature information that captures the user's features more accurately.

[7. Hardware configuration)
The learning device 100 and the user terminal 10 according to the above-described embodiment are realized by a computer 1000 having a configuration as shown in FIG. Hereinafter, the learning apparatus 100 will be described as an example. FIG. 17 is a hardware configuration diagram illustrating an example of a computer 1000 that implements the functions of the learning apparatus 100. The computer 1000 includes a CPU 1100, RAM 1200, ROM 1300, HDD 1400, communication interface (I / F) 1500, input / output interface (I / F) 1600, and media interface (I / F) 1700.

  The CPU 1100 operates based on a program stored in the ROM 1300 or the HDD 1400 and controls each unit. The ROM 1300 stores a boot program executed by the CPU 1100 when the computer 1000 is started up, a program depending on the hardware of the computer 1000, and the like.

  The HDD 1400 stores a program executed by the CPU 1100, data used by the program, and the like. The communication interface 1500 receives data from other devices via the communication network 500 (corresponding to the network N shown in FIG. 4), sends the data to the CPU 1100, and transmits the data generated by the CPU 1100 to the other devices via the communication network 500. Send to device.

  The CPU 1100 controls an output device such as a display and a printer and an input device such as a keyboard and a mouse via the input / output interface 1600. The CPU 1100 acquires data from the input device via the input / output interface 1600. Further, the CPU 1100 outputs the data generated via the input / output interface 1600 to the output device.

  The media interface 1700 reads a program or data stored in the recording medium 1800 and provides it to the CPU 1100 via the RAM 1200. The CPU 1100 loads such a program from the recording medium 1800 onto the RAM 1200 via the media interface 1700, and executes the loaded program. The recording medium 1800 is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) or PD (Phase change rewritable disk), a magneto-optical recording medium such as an MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. Etc.

  For example, when the computer 1000 functions as the learning apparatus 100 according to the embodiment, the CPU 1100 of the computer 1000 implements the function of the control unit 130 by executing a program loaded on the RAM 1200. The HDD 1400 stores data in the storage unit 120. The CPU 1100 of the computer 1000 reads these programs from the recording medium 1800 and executes them, but as another example, these programs may be acquired from other devices via the communication network 500.

[8. Others]
In addition, among the processes described in the above embodiment, all or part of the processes described as being automatically performed can be performed manually, or the processes described as being performed manually can be performed. All or a part can be automatically performed by a known method. In addition, the processing procedures, specific names, and information including various data and parameters shown in the document and drawings can be arbitrarily changed unless otherwise specified. For example, the various types of information illustrated in each drawing is not limited to the illustrated information.

  Further, each component of each illustrated apparatus is functionally conceptual, and does not necessarily need to be physically configured as illustrated. In other words, the specific form of distribution / integration of each device is not limited to that shown in the figure, and all or a part thereof may be functionally or physically distributed or arbitrarily distributed in arbitrary units according to various loads or usage conditions. Can be integrated and configured. For example, the acquisition unit 132 and the reception unit 136 illustrated in FIG. 5 may be integrated. Further, the specifying unit 137 illustrated in FIG. 5 calculates a probability that the user terminal 10 installs the application based on the model, and whether to distribute the advertisement to the user terminal 10 based on the calculated result. It may be distributed to a determination unit that determines whether the user terminal 10 is a distribution destination based on the determined result. Further, for example, information stored in the storage unit 120 may be stored in a predetermined storage device provided outside via the network N.

  Moreover, in the said embodiment, the learning apparatus 100 performs the reception process which receives the submission of an advertisement (information content), the learning process which learns the relationship between an application and a user, and the delivery process which distributes an advertisement, for example. An example to do is shown. However, the learning device 100 described above may be separated into a reception device that performs reception processing, a learning device that performs learning processing, and a distribution device that performs distribution processing. In this case, the reception device has at least a submission reception unit 131. The learning device has at least a learning unit 133. The distribution apparatus has at least a distribution unit 138. And the process by said learning apparatus 100 is implement | achieved by the learning processing system 1 which has each apparatus of a reception apparatus, a learning apparatus, and a delivery apparatus.

  In addition, the above-described embodiments and modifications can be combined as appropriate within a range that does not contradict processing contents.

[9. effect〕
As described above, the learning device 100 according to the embodiment includes the acquisition unit 132, the generation unit 135, and the specification unit 137. The acquisition unit 132 acquires terminal information that is information regarding the user terminal 10 that receives information content (for example, an advertisement). The generation unit 135 learns the relationship between the result information indicating whether or not the predetermined action related to the predetermined information content has been executed and the terminal information acquired by the acquisition unit 132, thereby determining the predetermined information related to the predetermined information content. Generate a model that shows the tendency of the action to be executed. Based on the model generated by the generation unit 135, the specification unit 137 specifies the user terminal 10 that is a providing destination that provides predetermined information content.

  As described above, the learning apparatus 100 according to the embodiment learns the relationship between the predetermined behavior for the information content to be provided and the terminal information of the provision destination to which the information content is provided, and thereby the model corresponding to the information content Can be generated. The learning apparatus 100 specifies the information content provision destination by using the tendency indicated by the generated model (for example, the probability that a predetermined action related to the predetermined information content is executed), thereby providing the information content to be provided and the provision destination It is possible to appropriately perform matching with other users.

  Further, the generation unit 135 learns the relationship between the result information and the feature information that is information extracted from the terminal information and that indicates the characteristics of the user terminal 10 or the user who uses the user terminal 10. To generate a model.

  As described above, the learning device 100 according to the embodiment extracts information indicating the user terminal 10 and the characteristics of the user of the user terminal 10 from the acquired terminal information. As a result, the learning device 100 can acquire the user terminal 10 and a large amount of information indicating the characteristics of the user of the user terminal 10, and thus more accurately reflect the characteristics of the information content providing destination in the model. it can. For this reason, the learning apparatus 100 can improve the accuracy of the model to be generated.

  Further, the generation unit 135 generates a model using attribute information that is information indicating a user attribute as feature information extracted from the terminal information. Specifically, the generation unit 135 generates a model using at least one of the user's sex, age, and residence as the attribute information.

  Thus, the learning apparatus 100 according to the embodiment generates a model using the user attribute information. The user attribute information includes a lot of information indicating the user's identity such as age and sex. That is, the learning apparatus 100 can generate a model that more reflects the user's characteristics by generating a model using attribute information. Thereby, the learning apparatus 100 can improve the accuracy of the model.

  Moreover, the acquisition part 132 acquires the apparatus information which is the information as a product of the user terminal 10 as terminal information. The generation unit 135 generates a model using feature information extracted from the device information. Specifically, the generation unit 135 has at least one of a model number, a brand name, a time elapsed since sales, a communication carrier name, a manufacturer name, and a resolution set as the feature information extracted from the device information. A model is generated using

  As described above, the learning device 100 according to the embodiment acquires information as a product of the user terminal 10 and uses the information as information indicating a feature as an information content providing destination. Since information as a product includes a brand, specifications, and the like, it can be an element that reflects characteristics such as preferences and personality of a user who uses the user terminal 10. And since the learning apparatus 100 can input a user's characteristic from various aspects by reflecting apparatus information in a model, the model which can specify the user who provides an information content accurately can be produced | generated. it can.

  In addition, the acquisition unit 132 acquires terminal information related to the user terminal 10 that can receive an advertisement related to the application and can install the application. The generation unit 135 learns the relationship between the result information indicating whether or not the predetermined application is installed in the user terminal 10 and the terminal information acquired by the acquisition unit 132, so that the predetermined application is installed. Generate a model that shows the tendency. Based on the model generated by the generation unit 135, the specification unit 137 specifies the user terminal 10 that is the destination of the advertisement related to the predetermined application.

  Thereby, the learning apparatus 100 according to the embodiment can generate a model for determining a user who can easily install an application associated with an advertisement. And learning device 100 distributes an advertisement to a user who tends to be similar to a user who installed an application in the past by specifying a distribution destination of an advertisement based on a model. For this reason, the learning apparatus 100 can appropriately perform matching between the distributed advertisement and the distribution destination user.

  Further, the acquisition unit 132 acquires application information that is information related to an application installed in the user terminal 10 as terminal information. The generation unit 135 generates a model indicating a tendency to install a predetermined application by learning the relationship between the result information and the feature information extracted from the application information.

  As described above, the learning device 100 according to the embodiment accurately estimates, for example, a user who tends to install a similar application by acquiring information about the application already installed in the user terminal 10. can do. Thereby, the learning device 100 can improve the accuracy of the model to be generated.

  Further, the generation unit 135 includes, as feature information extracted from the application information, the total number of applications installed in the user terminal 10, the number of game applications installed in the user terminal 10, and games other than the games installed in the user terminal 10. A model is generated using at least one of the number of apps or the ratio between the total number of apps installed in the user terminal 10 and the number of game apps.

  As described above, the learning device 100 according to the embodiment uses information regarding the number of applications installed in the user terminal 10 as feature information. For example, the user can install a new game app as the number of applications installed on the user terminal 10 increases as the number of apps installed on the user terminal 10 increases. It can be information indicating the behavior and characteristics of the user, such as being highly likely. The learning apparatus 100 can improve the accuracy of the model by learning such information and generating a model.

  Further, the generation unit 135 is information specified based on the type or element set in the application installed in the user terminal 10 as the feature information extracted from the application information. A model is generated using interest information, which is information to be shown. Further, the generation unit 135 is information specified based on the type or element set in the game application installed in the user terminal 10 as feature information extracted from the application information, and the game of the user of the user terminal 10 A model may be generated using preference information which is information indicating preference.

  As described above, the learning device 100 according to the embodiment extracts feature information that characterizes the user in consideration of not only the number of installed applications but also the type of application and elements of the application. For this reason, the learning apparatus 100 can perform detailed characterization regarding the user, and can improve the accuracy of the model by reflecting the information on the model.

  Further, the generation unit 135 sets the type or element of the application installed in the user terminal 10 based on the introduction of the application set for the application from the provider side that provides the application.

  As described above, the learning apparatus 100 according to the embodiment can set a type and an element for an application by analyzing, for example, an application introduction sentence registered in an application download site or the like. Thereby, the learning apparatus 100 can automatically set the type and the element for the application without artificially characterizing the type and the element for the application. For this reason, the learning apparatus 100 can reduce the effort of the administrator of the learning apparatus 100, for example.

  The embodiment of the present application has been described in detail with reference to the drawings. However, this is an exemplification, and various modifications and improvements are made based on the knowledge of those skilled in the art including the aspects described in the column of the disclosure of the invention. The present invention can be implemented in other forms.

  In addition, the “section (module, unit)” described above can be read as “means” or “circuit”. For example, the acquisition unit can be read as acquisition means or an acquisition circuit.

1 learning processing system 10 user terminal 20 advertiser terminal 30 web server 100 learning device 110 communication unit 120 storage unit 121 advertisement information storage unit 122 terminal information storage unit 123 attribute table 124 device table 125 application table 126 learning model storage unit 127 setting table 128 variable table 129 model table 130 control unit 131 submission acceptance unit 132 acquisition unit 133 learning unit 134 extraction unit 135 generation unit 136 reception unit 137 identification unit 138 distribution unit

Claims (14)

An acquisition unit that acquires application information that is information related to an application installed in the terminal device, as one of terminal information that is information related to the terminal device that receives information content associated with a predetermined application ;
And result information indicating whether or not a predetermined action concerning a given application associated with the predetermined information content is performed by learning the relationship between the terminal information obtained by the obtaining unit, the predetermined A generation unit that generates a model for each predetermined application indicating a tendency to execute a predetermined action related to the predetermined application associated with the information content;
Based on the model generated by the generating unit, a specifying unit that specifies a terminal device that provides the predetermined information content;
Learning and wherein the obtaining Bei a. The generator is
The model is obtained by learning the relationship between the result information and feature information that is information extracted from the terminal information and that indicates the characteristics of the terminal device or a user who uses the terminal device. Generate,
The learning apparatus according to claim 1. The generator is
As the feature information extracted from the terminal information, using the attribute information that is information indicating the attribute of the user, the model is generated.
The learning apparatus according to claim 2, wherein: The generator is
The model is generated using at least one of the user's sex, age, and residence as the attribute information.
The learning apparatus according to claim 3. The acquisition unit
As the terminal information, obtain device information that is information as a product of the terminal device,
The generator is
The feature information extracted from the device information is used to generate the model.
The learning device according to claim 2, wherein the learning device is a learning device. The generator is
The model is generated using at least one of the model number, brand name, time elapsed since sales, communication carrier name, manufacturer name, and resolution set as the terminal device as feature information extracted from the device information. ,
The learning apparatus according to claim 5. The acquisition unit
Get terminal information about a terminal device that can receive advertisements about apps, and that can install the app,
The generator is
A model that shows the tendency that the predetermined application is installed by learning the relationship between the result information indicating whether or not the predetermined application is installed in the terminal device and the terminal information acquired by the acquisition unit Produces
The specific part is:
Based on the model generated by the generation unit, the terminal device that is the destination of the advertisement related to the predetermined application is specified.
The learning apparatus according to claim 1, wherein The generator is
Learning a relationship between the result information and feature information extracted from the application information, thereby generating a model indicating a tendency to install the predetermined application;
The learning device according to claim 7. The generator is
As feature information extracted from the app information, the total number of apps installed in the terminal device, the number of game apps installed in the terminal device, the number of apps other than games installed in the terminal device, or Generating the model using at least one of a ratio between the total number of apps installed in the terminal device and the number of game apps;
The learning device according to claim 8. The generator is
The feature information extracted from the application information is information specified based on a type or element set in the application installed in the terminal device, and is information indicating interest of the user of the terminal device To generate the model,
The learning apparatus according to claim 8 or 9, wherein The generator is
The feature information extracted from the application information is information specified based on the type or element set in the game application installed in the terminal device, and information indicating the game preference of the user of the terminal device Generating the model using preference information;
The learning apparatus according to claim 8, wherein the learning apparatus is a learning apparatus. The generator is
Set the type or element of the application installed on the terminal device based on the introduction of the application set for the application from the provider side that provides the application.
The learning apparatus according to claim 10 or 11, wherein A learning method performed by a computer,
An acquisition step of acquiring application information that is information related to an application installed in the terminal device as one of terminal information that is information related to a terminal device that receives information content associated with a predetermined application ;
And result information indicating whether or not a predetermined action concerning a given application associated with the predetermined information content is performed by learning the relationship between the terminal information obtained by the obtaining step, the predetermined A generation step of generating a model indicating a tendency to execute a predetermined action related to a predetermined application associated with the information content for each predetermined application ;
Based on the model generated by the generating step, a specifying step for specifying a terminal device that provides the predetermined information content;
Learning wherein the containing octopus a. An acquisition procedure for acquiring application information that is information related to an application installed in the terminal device as one of terminal information that is information related to a terminal device that receives information content associated with a predetermined application ;
And result information indicating whether or not a predetermined action concerning a given application associated with the predetermined information content is performed by learning the relationship between the terminal information acquired by the acquisition procedure, the predetermined A generation procedure for generating, for each predetermined application, a model indicating a tendency to execute a predetermined action related to the predetermined application associated with the information content;
Based on the model generated by the generation procedure, a specific procedure for specifying a terminal device that provides the predetermined information content;
A learning program characterized by causing a computer to execute.

Images