JP6527276B2 - INFORMATION PROCESSING APPARATUS, INFORMATION PROCESSING PROGRAM, AND INFORMATION PROCESSING METHOD - Google Patents

Description

  The present invention relates to an information processing apparatus, an information processing program, and an information processing method.

  In recent years, along with the rapid spread of the Internet, advertisement delivery based on information acquired via a network has been performed. As an example of such advertisement distribution, there is known a technology for specifying an audience similar to an audience group having an attribute among users who browse web pages, and distributing the advertisement (for example, patent documents) 1).

JP 2012-94097 A

  However, Patent Document 1 described above only describes specifying an audience whose web browsing history is similar to an audience group targeted for advertisement distribution, and does not present a specific method.

  The present application has been made in view of the above, and it is an object of the present invention to provide an information processing apparatus, an information processing program, and an information processing method capable of effectively expanding a distribution target of content such as advertising content. .

  An information processing apparatus according to the present application includes: a model generation unit that generates a prediction model in which the presence or absence of a user's action history of goods and / or services sold in a predetermined e-commerce is an objective variable and the attribute of the user is an explanatory variable If the calculation result of the prediction model using an attribute of a specific user who is a user who has no action history of the product and / or the service as an explanatory variable is greater than or equal to a predetermined threshold, the action history of the specific user is determined to be present. And an updating unit for performing updating.

  According to one aspect of the embodiment, it is possible to effectively expand the distribution target user of the content such as the advertising content.

FIG. 1 is an explanatory diagram of user attribute update processing according to the embodiment. FIG. 2 is a diagram showing an example of the configuration of the advertisement distribution system according to the present embodiment. FIG. 3 is a diagram illustrating an example of the configuration of the information processing device 3 according to the embodiment. FIG. 4 is a diagram illustrating an example of the user information storage unit according to the embodiment. FIG. 5 is a diagram illustrating an example of coefficients and features of the prediction model according to the embodiment. FIG. 6 is a flowchart illustrating an example of the user attribute update processing procedure according to the embodiment. FIG. 7 is a hardware configuration diagram showing an example of a computer that realizes an information processing function according to the embodiment.

  Hereinafter, an information processing apparatus, an information processing program, and a mode for carrying out an information processing method according to the present application (hereinafter, referred to as “embodiment”) will be described in detail with reference to the drawings. Note that the information processing apparatus, the information processing program, and the information processing method according to the present application are not limited by the embodiment.

[1. User attribute update process]
First, user attribute update processing according to the embodiment will be described. FIG. 1 is an explanatory diagram of user attribute update processing according to the embodiment. In the following, user attribute update processing is performed by the information processing apparatus.

In the example illustrated in FIG. 1, the information processing apparatus generates a prediction model represented by the following Expression (1) as a prediction model used to update the attribute of the user. a _i is a coefficient, and x _i is a feature (explanatory variable). The information processing apparatus generates a prediction model by calculating a coefficient _ai by machine learning such as SVM (Support Vector Machine), for example.

Specifically, the information processing apparatus calculates the coefficient a _i of the above equation (1) using SVM with the information of the plurality of users U1 as correct answer data and the information of the plurality of users U2 as incorrect answer data, and prediction A model is generated (step S1). Here, the user U1 is, for example, a user who has a purchase history of a product sold in a predetermined e-commerce, and the user U2 is, for example, a user who has no purchase history of a product sold in a predetermined e-commerce.

In addition, as information of the users U1 and U2, for example, demographic attributes indicating attribute information of demographic users such as "sex" and "age", user preferences such as "car" and "fashion", values , Life style, psychographic attribute indicating personality, etc. The information processing apparatus generates a prediction model by calculating information of the users U1 and U2 as the feature x _i and using the purchase history as a dependent variable to calculate the coefficient a _i of the equation (1).

The information processing apparatus acquires one or more pieces of feature information based on the generated prediction model (step S2). Here, the feature information is information included in many users U1 among the information of the plurality of users U1 as correct data, and is information indicating the features of the plurality of users U1. Specifically, since the information included in the information about many users U1 corresponding to the coefficient a _i is greater feature x _i of the prediction model calculated in step S1, obtaining a coefficient a _i is greater feature x _i as feature information Do.

  Further, the information processing apparatus performs an operation based on the prediction model based on the information of the user U3 different from the user U1 (step S3). Here, the user U3 is, for example, a user who does not have a purchase history of a product sold in a predetermined e-commerce. The user U3 may be the same as or different from the user U2. If the result of the operation satisfies the predetermined condition, the information processing apparatus updates the attribute of the user U3 (step S4). Specifically, the feature information acquired in step 2 is set as attribute information of the user U3.

  As described above, in the user attribute update process according to the present embodiment, the information processing apparatus generates a prediction model by using information of the user U1 having a purchase history of a product sold in a predetermined e-commerce as correct data Based on the prediction model, the attributes of the user U3 different from the user U1 are updated. Further, the information processing apparatus extracts feature information based on the prediction model, and updates the attribute of the user U3 based on the extracted feature information.

  Therefore, for example, when distributing advertisement content to a user U1 who has a predetermined product purchase history, such advertisement content can be distributed to a user U3 which is different from the user U1 but similar to the user U1. Thus, for example, the user who distributes the advertisement content can be effectively expanded to the user U3 as well as the user U1 who was originally targeted for distribution.

  Hereinafter, an embodiment of an advertisement delivery system that includes the above-described information processing apparatus and delivers advertisement content as an example of content will be described.

[2. Advertising Delivery System]
FIG. 2 is a diagram for explaining an exemplary configuration of the advertisement distribution system 1 according to the present embodiment. As shown in FIG. 2, the advertisement distribution system 1 according to the present embodiment includes a web server 2, an information processing device 3, and a plurality of terminal devices 5. These devices are communicably connected to each other via the communication network 6. The communication network 6 is, for example, the Internet.

  The terminal device 5 is, for example, a personal computer (PC) used by the user U, a personal digital assistant (PDA), a tablet terminal, a smartphone, or the like. For example, a browser application (hereinafter referred to as a browser) is installed in the terminal device 5.

  The web server 2 stores a plurality of web pages in which advertisement spaces are set. When accessed by the browser of the terminal device 5 via the communication network 6, the control unit of the web server 2 provides a web page corresponding to a URL (Uniform Resource Locator) designated by the terminal device 5.

  When the browser of the terminal device 5 receives the web page from the web server 2, the browser of the terminal device 5 transmits an advertisement request corresponding to the advertisement space set in the web page to the information processing device 3. The advertisement request is a distribution request for advertisement content to be displayed in the advertisement space, and, for example, identification information of the user U of the terminal device 5 (hereinafter referred to as user ID) or identification information of the advertisement space (hereinafter referred to as advertisement space ID To include). The user ID is, for example, an HTTP (Hypertext Transfer Protocol) cookie or the like.

  The information processing device 3 distributes the advertisement content corresponding to the advertisement request to the terminal device 5. The information processing device 3 distributes the advertising content based on the attribute of the user U. Note that such advertisement content is, for example, a banner advertisement, and when clicked by the user U, causes transition to the web page of the advertiser.

  Hereinafter, the configuration example of the information processing device 3 will be described in more detail.

[2.1. Information processing device 3]
FIG. 3 is a view showing a configuration example of the information processing device 3. As shown in FIG. 3, the information processing device 3 includes a communication unit 10, a control unit 20, and a storage unit 30.

  The communication unit 10 is a communication interface that transmits and receives information to and from the communication network 6, and performs connection with the communication network 6 by wire or wirelessly. The control unit 20 can transmit and receive various types of information to and from the terminal device 5 and other devices via the communication unit 10 and the communication network 6.

  The storage unit 30 includes a user information storage unit 31 and an advertisement information storage unit 32. The user information storage unit 31 and the advertisement information storage unit 32 are, for example, a storage device such as a random access memory (RAM), a semiconductor memory device such as a flash memory, or a hard disk or an optical disk.

[2.1.1. User Information Storage Unit 31]
Here, FIG. 4 shows an example of the user information storage unit 31 according to the present embodiment. FIG. 4 is a diagram showing an example of the user information storage unit 31 according to the embodiment. The user information storage unit 31 stores a purchase history of a product in a predetermined e-commerce of the user U and attribute information of the user U.

  Here, the predetermined e-commerce is electronic commerce performed via the communication network 6 such as auction or shopping. Hereinafter, it will be described that the purchase history of the product in the predetermined e-commerce is a successful bid history of the auction.

  In FIG. 4, the user information storage unit 31 has items such as “user ID”, “demographic attribute”, “psychographic attribute”, and “successful bid history”.

  The “user ID” is identification information for identifying the user U of the terminal device 5. In FIG. 4, the user ID is written as "U11" or the like. This indicates that the terminal device 5 is identified by the user ID "U11". Here, it is assumed that the user ID matches the reference code of the user who operates the terminal device 5. Next, “demographic attribute” shown in FIG. 4 is divided into types such as “sex” and “age”. Thus, the “demographic attribute” indicates attribute information of the demographic user U.

  The “psychographic attribute” indicates the user's preference, values, lifestyle, character and the like. In the example illustrated in FIG. 4, the user information storage unit 31 stores, as the “psychographic attribute”, a feature word related to an object of interest of the user U. Specifically, the user information storage unit 31 uses, as feature words, keywords indicating the interest of the user U, such as "car", "fashion", "travel", "sports", etc., as the "psychographic attribute". Remember.

  The characteristic word is a word characteristically expressing the interest of the user U, and is a word showing the characteristic of the user U. Thus, the feature word is a type of feature information that is information indicating the feature of the user U. The feature information of the user U includes, for example, the interest degree of the user U for a predetermined keyword besides the feature word. . The characteristic word may be, for example, a keyword extracted from a web page browsed by the user U, or may be a keyword extracted from a search query input to the search web page by the user U.

  The “successful bid history” stores information of goods for which the user U made a successful bid on the auction site. In the example shown in FIG. 4, the user information storage unit 31 associates and stores the product name for which a successful bid is made and the category of the product as “the successful bid history”. For example, FIG. 4 shows an example in which the user U11 purchases a product "truck" whose major category is "model" and whose minor category is "car". In addition, the users U31 and U32 having a blank successful bid history indicate that they have never made a successful bid for a product at the auction site.

[2.1.2. Control unit 20]
The control unit 20 illustrated in FIG. 3 is realized by, for example, an integrated circuit such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA). The control unit 20 executes a program stored in an internal storage device by an internal central processing unit (CPU) or a micro processing unit (MPU) using a RAM as a work area, whereby the model generation unit 21 and the acquisition unit 22 are obtained. Functions as the calculating unit 23, the updating unit 24, and the advertisement distributing unit 25. The configuration of the control unit 20 is not limited to this configuration, and may be another configuration as long as it performs the information processing described later.

[2.1.3. Model generation unit 21]
The model generation unit 21 acquires, from the user information storage unit 31, information of the user U who has a successful bid history of a product on the auction site among the users U. Further, the model generation unit 21 acquires, from the user information storage unit 31, information of the user U having no successful bid history of the product among the users U. The model generation unit 21 generates information on a user U having a successful bid history of a product as correct solution data, generates information on the user U having no successful bid history on the product as incorrect solution data, and stores the generated predictive model as a storage unit. Remember to 30.

Hereinafter, the process of generating a prediction model according to the embodiment will be specifically described using FIGS. 4 and 5. FIG. 5 is a diagram showing the relationship between the coefficient a _{i of the} prediction model generated by the model generation unit 21 and the feature x _i . Here, the model generation unit 21 corrects the user information of the users U11, U12,... (Hereinafter also referred to as the user U1) having a successful bid history of a product whose major category is “model” and the minor category is “car”. It is data, and there is a successful bid history of a product with a large category "model" but user information of users U21, U22, ... (hereinafter also referred to as user U2) without a successful bid history of a product with a small category "car". An example will be described in which a prediction model is generated with as incorrect answer data.

  Here, user U1 with a successful bid history of a product with a large category "model" and a small category "car" is regarded as correct data, and there is a successful bid history of a product with a large category "model", but the small category is "car" The reason for making the user U2 having no successful bid history of the item of <1> as incorrect answer data will be described. Since the user U1 has a successful bid history of products having a large category "model" and a small category "car", it is considered that the degree of interest in "car" among the "model" is relatively high.

  On the other hand, since user U2 makes a successful bid for a product whose large category is included in "model", it is considered that "model" is interested, and there is no successful bid history for products whose small category is "car" Thus, it is considered that the user U2 has a relatively low degree of interest in the "car" among the "model".

  Thus, the user U1 and the user U2 have different degrees of interest in the "car" of the "model". Therefore, the model generation unit 21 according to the embodiment predicts the user U who is interested in the "car" of the "model" by generating the prediction model with the user U1 as the correct answer data and the user U2 as the incorrect answer data. Can generate a prediction model.

  Here, although an example in which the model generation unit 21 generates a prediction model paying attention to the “car” which is a small category will be described, the present invention is not limited thereto. For example, when generating a prediction model for predicting a user U who is interested in a specific product, the model generation unit 21 sets the user U having a successful bid history of the specific product as correct data and does not have a successful bid history of the product. A prediction model can also be generated with the user U as incorrect answer data. Further, the prediction model can be similarly generated even in accordance with the presence or absence of the successful bid history of the products included in the large category.

For example, in the SVM shown in the above-mentioned formula (1), the model generation unit 21 uses, for example, the presence or absence of the purchase of the product included in the above-mentioned small category "car" as a dependent variable, and makes the above user information a feature (explanatory variable) x _By setting _i , the coefficient a _i corresponding to the feature x _i is obtained.

The feature x _i is, for example, an attribute of the user U. For example, the feature x _i is a feature word stored in the user information storage unit 31 as “psychographic attribute”. Specifically, for example, the feature x ₁ is “car”, and is set to “1” when “car” is stored in the user information storage unit 31 as the “Psychographic attribute” of the user U. The others are set to "0". Also, the feature x ₂ is “child”, and is set to “1” when “child” is stored in the user information storage unit 31 as the “Psychographic attribute” of the user U, and “0” is otherwise Is set to Thus, the attribute information stored in the user information storage section 31 as a "psychographic attribute" is assigned to the feature x _i.

  Further, the dependent variable is set to “1”, for example, when the user U makes a successful bid for a product included in the large category “model” and the small category “car” in the successful bid history, and the user U sets the large category “model And is set to “−1” when a product for which the item is included and not included in the small category “car” is awarded. Specifically, the users U11, U12,... Of FIG. 4 are set to “1”, and U21, U22,.

Thus, the model generation unit 21 generates a prediction model based on SVM. The model generation unit 21 associates the feature x _{i of the} prediction model generated as shown in FIG. 5 with the coefficient a _i and stores the result in the storage unit 30. The above-described dependent variable and feature (explanatory variable) are an example, and the model generation unit 21 can also generate a prediction model using other information or part of the above-described information. For example, as a dependent variable, "-1" may be set when a product is awarded, and "1" may be set when the product is not awarded.

The model generation unit 21 can also generate a prediction model using a machine learning method such as, for example, a logistic regression model. For example, when generating a prediction model using a logistic regression model, the model generation unit 21 sets the presence or absence of purchase of a product as a dependent variable in the following equation (2) and sets the user information described above to the feature (explanatory variable) x _i Then, the coefficient a _i corresponding to the feature x _i may be obtained.

[2.1.4. Acquisition unit 22]
The acquisition unit 22 in FIG. 3 is an extraction and acquisition unit that extracts and acquires one or more pieces of feature information based on the prediction model generated by the model generation unit 21. Specifically acquisition unit 22, among the coefficients a _i of the formula (1) calculated by the model generating unit 21, extracts the N pieces of feature x _i corresponding to N coefficients in order of large value.

When, for example, the coefficient _ai shown in FIG. 5 is calculated by the model generation unit 21, the acquisition unit 22 extracts and acquires three pieces of feature information. In this case, the acquisition unit 22 acquires three feature words of “a toy”, “a car”, and “a child” as feature information. The acquisition unit 22 stores the extracted and acquired feature information in the storage unit 30.

  In addition, although the case where the acquisition part 22 extracts feature information based on a prediction model was demonstrated here, it is not restricted to this. For example, the storage unit 30 may store the product or category and the feature information in advance in association with each other, and the acquisition unit 22 may acquire the feature information from the storage unit 30.

  In addition, an example is shown in which the same prediction model is used for extraction and acquisition of feature information by the acquisition unit 22 and update of a user attribute by the update unit 24. Here, the prediction model is an example, and the acquisition unit 22 can also extract and acquire feature information based on a prediction model different from the prediction model used for updating the user attribute by the update unit 24.

  Further, the number of pieces of feature information extracted by the acquisition unit 22 may be changed according to categories, products, and the like. Alternatively, the acquisition unit 22 may change the number of pieces of feature information to be extracted according to the calculation result of the prediction model.

[2.1.5. Arithmetic unit 23]
The calculation unit 23 of FIG. 3 is generated by the model generation unit 21 based on the information of the user U different from the user U1, that is, the user U having no purchase history of products included in the large category "model" and the small category "car". Perform calculations using a prediction model. For example, in the example shown in FIG. 4, the users U31, U32,... (Hereinafter also referred to as user U3) having no purchase history at the auction site, and the products included in the large category "model" and the small category "car" Based on the information of the user U2 having no purchase history, the calculation by the prediction model is performed.

  As described above, the calculation unit 23 performs the calculation based on the prediction model based on the information of the user U including the user U2 as the incorrect answer data. By this, it is predicted whether the user U2 who thinks that the "car" of the "model" is not interested has the user U2 which may have an interest of interest in the "car" of the "model". it can.

For example, when performing an operation based on the user information of the user U31 illustrated in FIG. 4, the operation unit 23 acquires the user information of the user U31 from the user information storage unit 31. The operation unit 23 performs the operation based on the prediction model of Expression (1) with the “psychographic attribute” acquired from the user information storage unit 31 as the feature x _i .

Specifically, for example, as shown in FIG. 4, the user U31 has a psychographic attribute of "child". As shown in FIG. 5, “child” is feature x ₂ . Therefore, calculation unit 23 sets the feature _{x 2} to "1". Thus, the attribute information stored in the user information storage section 31 as a "psychographic attribute" is assigned to the feature x _i. The operation unit 23 performs an operation according to the prediction model by multiplying the allocated feature x _i by the coefficient a _i to obtain the sum. The operation unit 23 stores the user U and the operation result in the storage unit 30 in association with each other.

[2.1.6. Update unit 24]
The update unit 24 updates the attributes of the users U2 and U3 for which the calculation unit 23 has performed the calculation using the prediction model. The update unit 24 updates the attribute information of the users U2 and U3 when the calculation result by the calculation unit 23 satisfies a predetermined condition.

  Here, the updating unit 24 updates the attribute of the user U when the number of the user U1 determined as correct data satisfies the predetermined condition. For example, when the number of users U1 is equal to or less than a predetermined threshold, the updating unit 24 of the user U2 who has purchased the item of the small category "railroad" included in the large category "model" including the item purchased by the user U1. Decide to update the attribute.

  In addition, when the number of users U2 is equal to or less than a predetermined threshold, the updating unit 24 may determine to update the attribute of the user U3 having no purchase history at the auction site, for example. As described above, the updating unit 24 may also add the user who updates the attribute according to, for example, the presence or absence of the purchase of the product included in the category until the number of the user U who updates the attribute exceeds a predetermined threshold. it can.

  The updating unit 24 notifies the computing unit 23 of the user U determined to update the attribute. The calculating unit 23 that has received the notification performs calculation using the prediction model based on the information of the user U who has determined to update the attribute. The update unit 24 updates the attribute of the user U according to the calculation result of the calculation unit 23.

  Specifically, the update unit 24 adds the feature information acquired by the acquisition unit 22 to the user information when the calculation result by the calculation unit 23 is equal to or more than a predetermined threshold. The updating unit 24 newly stores the feature words acquired by the acquiring unit 22 in the user information storage unit 31 in association with the users U2 and U3. Specifically, the feature word extracted as feature information is added to the psychographic attribute shown in FIG. 4 as a new item. For example, it is assumed that the user attributes of the user U31 are updated by adding the characteristic words "toe", "car", and "child" as feature information to the psychographic attribute of the user U31 in FIG. In this case, since the user U31 in FIG. 4 already has the feature word "child", the updating unit 24 adds "toy" and "car" as psychographic attributes to the user information of the user U31.

  Here, although the example in which the updating unit 24 updates the attribute information of the users U2 and U3 has been described, the present invention is not limited thereto. For example, the updating unit 24 may update the successful bid history. Specifically, the updating unit 24 adds information for making a successful bid for a product of a predetermined category to the successful bid history of the user U whose calculation result is equal to or more than a predetermined threshold value. In this case, the updating unit 24 adds, for example, a "model" to the large category of the successful bid history of the user U in the user information storage unit 31, and a "car" to the small category. At this time, the product name may be left blank, or a temporary product name may be added on the assumption that the temporary product is made a successful bid.

  Further, when the user information storage unit 31 digitizes and stores the successful bid history of the user U (for example, when the product is made a successful bid, it is set as “1”, and when not making a successful bid, it is stored as “0”). Alternatively, the successful bid history of the user U whose calculation result is equal to or more than a predetermined threshold may be updated from “0” to “1”. Note that such numerical values are not limited to “0” and “1”. For example, the user U who has made a successful bid for a product may be "2", and the user U whose calculation result is equal to or greater than a predetermined threshold may be "1". In addition, the user U whose calculation result is equal to or more than the predetermined threshold may store the calculation result as the successful bid history of the user U.

  Further, the example has been described in which the updating unit 24 updates the user attribute by adding the feature information acquired by the acquiring unit 22 to the “psychographic attribute” of the user information. Here, for example, the updating unit 24 may store the feature information and attribute information corresponding to at least one of an auction, a category, or a product in the user information storage unit 31 in association with each other. Alternatively, the information processing device 3 may store the feature information extracted and acquired by the prediction model separately from the “psychographic attribute” acquired directly from the user U. In this case, for example, when the psychographic attribute is to be overwritten or deleted, the feature information extracted and acquired by the prediction model may be preferentially overwritten or deleted.

  Note that the updating unit 24 can also periodically update the demographic attribute or the psychographic attribute described above based on, for example, the web browsing history or web search history of the user U, the input information by the user U, and the like. Further, the updating unit 24 can also periodically acquire and update, for example, the above-described demographic attribute or psychographic attribute from an external server.

  Here, although the updating unit 24 determines the user U whose attribute is to be updated, the present invention is not limited to this. For example, the operation unit 23 may determine the user U whose attribute is to be updated. Alternatively, the control unit 20 may include a determination unit (not shown) that determines the user U whose attribute is to be updated.

[2.1.7. Advertisement distribution unit 25]
When the advertisement delivery unit 25 of FIG. 3 receives a request for advertisement delivery from the terminal device 5 of the user U, the advertisement delivery unit 25 extracts advertisement content suitable for the user U from the advertisement information storage unit 32. The advertisement distribution unit 25 distributes the advertisement content extracted through the communication unit 10 to the terminal device 5. The advertisement distribution unit 25 uses the user information on the user U stored in the user information storage unit 31 to execute an appropriate matching process between the user U and the advertisement content. In other words, the advertisement distribution unit 25 extracts appropriate advertisement content corresponding to the specific user U whose appealing effect is expected to be high, and distributes the extracted advertisement content to the user U. At this time, the distribution target user of the content can be effectively expanded as the user information by the update by the update unit 24, and the appeal effect of the content to the user U can be enhanced.

[3. Processing procedure]
Next, the procedure of the user attribute update process performed by the information processing device 3 according to the embodiment will be described with reference to FIG. FIG. 6 is a flowchart showing a user attribute update processing procedure by the information processing apparatus 3 according to the embodiment.

  As shown in FIG. 6, the information processing apparatus 3 acquires information on the user U1 having a successful bid history of a product included in a predetermined category of an auction as correct answer data, and there is no successful bid history of a product included in the predetermined category. Information of the user U2 is acquired as incorrect answer data (step S101).

  Subsequently, the information processing device 3 generates a prediction model using the correct answer data and the incorrect answer data (step S102). The information processing device 3 extracts and acquires feature information based on the prediction model (step S103).

  The information processing device 3 selects one of the users U2 and U3 who have not purchased the product purchased by the user U1 as the correct data, as the user to be calculated by the prediction model (step S104). The information processing device 3 performs an operation using a prediction model based on the information of the user U selected in step S104 (step S105). The information processing device 3 compares the calculation result with a predetermined threshold Th (step S106). If the calculation result is smaller than the predetermined threshold Th (step S106; No), the process proceeds to step S108.

  On the other hand, if the calculation result is equal to or more than the predetermined threshold Th (step S106; Yes), the user attribute selected in step S104 is updated based on the feature information extracted and acquired in step S103 (step S107). The information processing device 3 determines whether or not the calculation based on the prediction model has been performed on all the users U2 and U3 who have not purchased the product purchased by the user U1 as the correct data (step S108). If it is determined that the operation has been performed on all of the users U2 and U3 (step S108; Yes), the process ends. On the other hand, when there is a user U2 or U3 who has not performed an operation (Step S108; No), the process returns to Step S104, and an operation based on a prediction model is performed on the user U2 or U3 who has not performed an operation.

  Note that FIG. 6 illustrates an example in which the information processing device 3 extracts and acquires feature information and then performs an operation using a prediction model. However, the present invention is not limited to this example, and the information processing device 3 may extract and acquire feature information after performing an operation using a prediction model. In this case, the number of feature information to be extracted and acquired may be changed according to the calculation result by the prediction model. For example, when the value of the calculation result by the prediction model is large, the number of pieces of feature information to be acquired may be increased, and when the value is small, the number may be reduced.

  In addition, as described above, when generating a prediction model in the model generation unit 21, "-1" is set when a product is sold as a dependent variable, and "1" is set when a product is not sold. If it is set, the magnitude relationship between the calculation result in step S106 and the predetermined threshold value Th is reversed. That is, when the calculation result is larger than the predetermined threshold Th in step S106, the process proceeds to step S108, and when the calculation result is equal to or less than the predetermined threshold Th, the process proceeds to step S107.

[4. Modified example]
The embodiment described above may be implemented in various different forms other than the above embodiment. Therefore, another embodiment will be described below.

  Further, in the above embodiment, although the purchase history in the predetermined e-commerce has been described as the successful bid history in the auction, the present invention is not limited to this. For example, it is determined whether the information of the user U is to be the correct data or the incorrect solution data as the shopping history of shopping, the bidding history at the auction, the registration history to the favorite, etc. as the purchasing history at the predetermined e-commerce. May be The purchase history of the user U is not limited to the product. It may be a service, or it may be a service and a product.

  As described above, the information processing apparatus 3 determines that the information of the user U1 having the purchase history of the product and / or service sold in the predetermined e-commerce is correct data, and the user U2 does not have the purchase history of the product and / or service. , U3 information as incorrect answer data. Thereby, the prediction accuracy of the prediction model can be improved, the distribution target user of the content can be effectively expanded, and the appealing effect of the content can be enhanced.

Also, an example has been shown in which the user information storage unit 31 stores the feature word itself as the "psychographic attribute". Here, the user information storage unit 31 sets the feature word as an item of “psychographic attribute”, and sets “1” or “0” to the item corresponding to the feature word to set the feature word and the user U You may make it match and memorize | store. In this case, when the updating unit 24 updates the user attribute based on the feature word acquired by the acquiring unit 22, “1” may be set to the item corresponding to the feature word. Alternatively, a predetermined numerical value may be set for each item according to the degree of interest of the user U. In this case, the predetermined numerical value corresponding to the degree of interest may be determined based on, for example, the coefficient a _{i of the} prediction model.

  In the above embodiment, the user information storage unit 31 of the information processing device 3 stores the attribute information of the user U and the purchase history of the product. Here, the information processing device 3 may store the attribute information of the user U and the purchase history of the product in different storage units. In this case, the product purchase history can also be stored, for example, as a log of the user U's action history.

  Further, among the processes described in the above embodiment, all or part of the process described as being automatically performed may be manually performed, or the process described as being manually performed. All or part of them can be performed automatically by known methods. In addition, information including processing procedures, specific names, various data and parameters shown in the above-mentioned documents and drawings can be arbitrarily changed unless otherwise specified. For example, the various information shown in each figure is not limited to the illustrated information.

  Further, each component of each device illustrated is functionally conceptual, and does not necessarily have to be physically configured as illustrated. That is, the specific form of the distribution and integration of each device is not limited to the illustrated one, and all or a part thereof may be functionally or physically dispersed in any unit depending on various loads, usage conditions, etc. It can be integrated and configured.

  For example, the advertisement information storage unit 32 illustrated in FIG. 3 may be held by a storage server or the like (not shown) without being held by the information processing device 3. In this case, the information processing device 3 acquires, for example, information related to an advertisement from the storage server. Further, for example, the information processing device 3 includes a user information management device including a model generation unit 21, an acquisition unit 22, an operation unit 23, an update unit 24 and a user information storage unit 31, an advertisement delivery unit 25 and an advertisement information storage unit And an advertisement delivery device having

  Moreover, in the said embodiment, the advertisement delivery system 1 which delivers advertisement content as an example of content was demonstrated. Here, the content may be content other than advertising content, such as article content or moving image content.

[5. Hardware configuration]
Note that the information processing device 3 in the above-described embodiment is realized by, for example, a computer 100 configured as shown in FIG. 7. FIG. 7 is a hardware configuration diagram showing an example of the computer 100 for realizing the information processing function. The computer 100 includes a CPU 301, a RAM 302, a read only memory (ROM) 303, a hard disk drive (HDD) 304, a communication interface (I / F) 305, an input / output interface (I / F) 306, and a media interface (I / F). ) 307.

  The CPU 301 operates based on a program stored in the ROM 303 or the HDD 304 to control each part. The ROM 303 stores a boot program executed by the CPU 301 when the computer 100 starts up, a program depending on the hardware of the computer 100, and the like.

  The HDD 304 stores a program executed by the CPU 301, data used by the program, and the like. The communication interface 305 receives data from another device via the communication line 309 and sends the data to the CPU 301, and transmits data generated by the CPU 301 to the other device via the communication line 309.

  The CPU 301 controls an output device such as a display or a printer and an input device such as a keyboard or a mouse via the input / output interface 306. The CPU 301 acquires data from an input device via the input / output interface 306. The CPU 301 also outputs the generated data to an output device via the input / output interface 306.

  The media interface 307 reads a program or data stored in the recording medium 308 and provides the CPU 301 via the RAM 302. The CPU 301 loads the program from the recording medium 308 onto the RAM 302 via the media interface 307, and executes the loaded program. The recording medium 308 is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) or a PD (Phase change rewritable Disk), a magneto-optical recording medium such as a MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. Etc.

  When the computer 100 functions as the information processing device 3 in the embodiment, the CPU 301 of the computer 100 executes the program loaded on the RAM 302 to execute the model generation unit 21, the acquisition unit 22, and the calculation unit 23 of the control unit 20. , And the functions of the updating unit 24 and the advertisement distribution unit 25 are realized. In addition, data in the storage unit 30 is stored in the HDD 304.

  The CPU 301 of the computer 100 reads these programs from the recording medium 308 and executes them, but as another example, these programs may be acquired from another device via the communication line 309.

[6. effect]
As described above, the information processing device 3 according to the embodiment generates a prediction model by using information of the first user (user U1) who has a purchase history of goods and / or services sold by predetermined e-commerce as correct data. Based on the prediction model, the attributes of the second user (user U2, U3) different from the first user are updated. Thereby, the information processing device 3 can effectively extend the distribution target user of the content to the second user as well as the first user, and can enhance the appeal effect of the content.

  In addition, the updating unit 24 of the information processing device 3 of the embodiment updates the attribute of the second user determined that the calculation result by the prediction model satisfies the predetermined condition based on the information of the second user. As a result, the information processing device 3 can effectively extend the distribution target user of the content to the second user whose calculation result by the prediction model satisfies the predetermined condition, and can enhance the appeal effect of the content.

  Further, the updating unit 24 of the information processing device 3 according to the embodiment determines at least one of predetermined e-commerce, a product, and a service among the attributes of the second users U2 and U3 based on the prediction model generated by the model generation unit 21. Update the attribute corresponding to one. Thereby, the information processing device 3 can effectively expand the distribution target user of the content, and at the same time, update the attribute corresponding to at least one of the predetermined e-commerce, the product, and the service to appeal the content. The effect can be enhanced.

  Further, the updating unit 24 of the information processing device 3 of the embodiment updates the attribute information of the user who does not have the purchase history of the product and / or service in the predetermined e-commerce based on the prediction model. Thereby, the information processing device 3 can effectively expand the distribution target user of the content.

  Further, the acquiring unit 22 of the information processing device 3 according to the embodiment extracts and acquires one or more pieces of feature information based on the prediction model, and the updating unit 24 updates the attributes of the users U2 and U3 based on the feature information. . As a result, the content can be distributed to the users U2 and U3 based on the characteristic information, the distribution target user of the content can be effectively expanded, and the appealing effect of the content can be enhanced.

  Further, the model generation unit 21 of the information processing device 3 of the embodiment generates, as a prediction model used by the acquisition unit 22, a prediction model different from the prediction model used by the update unit 24. Thus, even when the prediction model is different, the feature information can be extracted, and the distribution target user of the content can be effectively expanded.

  In addition, the information processing device 3 according to the embodiment distributes advertising content based on the attributes of the users U2 and U3 updated by the updating unit 24. As a result, the distribution target user of the content can be effectively expanded, and the content with high appealing effect can be distributed to the expanded user.

  In addition, the updating unit 24 of the information processing device 3 according to the embodiment includes attribute information of a second user who has a purchase history of a product and / or service of a category including the product and / or service for which the first user has purchased history. Update based on the prediction model. Thereby, the distribution target user of the content can be effectively extended to the second user who has the purchase history of the goods and / or services of the category including the goods and / or services for which the first user has made the purchase history.

  Further, the updating unit 24 of the information processing device 3 of the embodiment updates the attribute of the second user when the number of first users satisfies the predetermined condition. As a result, the distribution target users of the content can be effectively expanded according to the number of first users.

  Although some of the embodiments of the present application have been described above in detail based on the drawings, these are only examples, and various modifications can be made based on the knowledge of those skilled in the art, including the aspects described in the section of the description of the invention. It is possible to implement the invention in other improved forms.

  In addition, the "section (module, unit)" described above can be read as "means" or "circuit". For example, the computing unit can be read as computing means or a computing circuit.

DESCRIPTION OF SYMBOLS 1 advertisement distribution system 2 web server 3 information processing apparatus 5 terminal device 6 communication network 10 communication unit 20 control unit 21 model generation unit 22 acquisition unit 23 calculation unit 24 update unit 25 advertisement distribution unit 30 storage unit 31 user information storage unit 32 advertisement Information storage unit

Claims (9)

A model generation unit that generates a prediction model in which the presence or absence of a user's action history of goods and / or services sold in a predetermined e-commerce is an objective variable, and the attribute of the user is an explanatory variable;
The action history of the specific user is determined to be present when the calculation result of the prediction model using an attribute of a specific user who is a user who has no action history of the product and / or the service as an explanatory variable is equal to or greater than a predetermined threshold. An updating unit that performs updating,
An information processing apparatus comprising: The updating unit is
An attribute corresponding to at least one of the predetermined e-commerce, the product, and the service among the attributes of the specific user as the characteristic information for updating the attribute of the specific user with an explanatory variable having a large coefficient in the prediction model The information processing apparatus according to claim 1, wherein the information processing apparatus is updated. And an extraction and acquisition unit for extracting and acquiring the feature information based on the prediction model generated by the model generation unit.
The updating unit is
The information processing apparatus according to claim 2, wherein the attribute of the specific user is updated based on the feature information. The model generation unit
The information processing apparatus according to claim 3, wherein a prediction model different from the prediction model used by the update unit is generated as the prediction model used by the extraction acquisition unit. The updating unit is
The attribute of the specific user who has a purchase history of a product and / or service of a category including the product and / or the service is updated based on the feature information. The information processing apparatus described in one. The updating unit is
The information processing apparatus according to any one of claims 1 to 5, wherein when the number of users having the action history satisfies a predetermined condition, an attribute of the specific user is updated. The information processing apparatus according to any one of claims 1 to 6, further comprising a distribution unit configured to distribute content based on the attribute of the specific user updated by the update unit. An information processing method executed by a computer,
Generating a prediction model in which the presence or absence of a user's action history of goods and / or services sold in a predetermined e-commerce is an objective variable, and the attribute of the user is an explanatory variable;
The action history of the specific user is determined to be present when the calculation result of the prediction model using an attribute of a specific user who is a user who has no action history of the product and / or the service as an explanatory variable is equal to or greater than a predetermined threshold. The process of updating,
An information processing method comprising: On the computer
A procedure for generating a prediction model in which the presence or absence of a user's action history of goods and / or services sold in a predetermined e-commerce is an objective variable, and the attribute of the user is an explanatory variable;
Update for setting the action history of the specific user when the calculation result of the prediction model using an attribute of a specific user who is a user who has no action history of the product and / or the service as an explanatory variable is equal to or more than a predetermined threshold The steps to do
An information processing program characterized by performing.

Images