Information processing device and information processing method

Abstract

To make it possible to accurately estimate an explanatory variable important for an objective variable.SOLUTION: An information processing device according to the present application includes: a generation unit which generates, on the basis of a plurality of pieces of set data being a set of explanatory data corresponding to each of a plurality of types of explanatory variables and objective data corresponding to objective variables, a first decision tree model for predicting objective data from first explanatory data corresponding to each of a first type of explanatory variables among the plurality of types of explanatory variables, and generates a second decision tree model for predicting objective data from second explanatory data corresponding to a second type of explanatory variable different from the first type of explanatory variable of a plurality of types of explanatory variables on the basis of set data corresponding to each of leaf nodes of the generated first decision tree; and an acquisition unit which acquires a second variable importance level generated when the generation unit has generated the second decision tree.SELECTED DRAWING: Figure 7

Description

Technical Field

[0001] The present invention relates to an information processing apparatus and an information processing method.

Background Art

[0002] Conventionally, a technique for estimating explanatory variables important for a target variable has been known. For example, a first value indicating the importance of each of a plurality of explanatory variables for a predetermined target variable is calculated using each of a plurality of m different methods. And a technique for selecting n explanatory variables based on a plurality of first values corresponding to each of the m different methods is known.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in the above-described conventional technique, since it only calculates a first value indicating the importance of each of a plurality of explanatory variables for a predetermined target variable using a plurality of different methods, it is not always possible to accurately estimate explanatory variables important for the target variable.

[0005] An object of the present application is to enable accurate estimation of explanatory variables important for a target variable.

Means for Solving the Problems

[0006] The information processing apparatus according to the present application generates a first decision tree model that predicts the target data from first explanatory data corresponding to each of the first type of explanatory variables among the plurality of types of explanatory variables, based on a plurality of set data that are sets of explanatory data corresponding to each of the plurality of types of explanatory variables and target data corresponding to a target variable. Then, based on the set data corresponding to each leaf node of the generated first decision tree model, a second decision tree model that predicts the target data from second explanatory data corresponding to a second type of explanatory variable different from the first type of explanatory variable among the plurality of types of explanatory variables is generated. The apparatus further includes an acquisition unit that acquires a second variable importance generated when the generation unit generates the second decision tree model.

Advantages of the Invention

[0007] According to one aspect of the embodiment, it is possible to accurately estimate explanatory variables important for a target variable.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Embodiments for Carrying Out the Invention

[0009] Hereinafter, embodiments for carrying out the information processing apparatus and the information processing method according to the present application (hereinafter referred to as "embodiments") will be described in detail with reference to the drawings. Note that the information processing apparatus and the information processing method according to the present application are not limited by this embodiment. Also, in the following embodiments, the same parts are denoted by the same reference numerals, and duplicate explanations are omitted.

[0010] (Embodiment) [1. Introduction] FIG. 1 is a diagram for explaining NPS (registered trademark) (Net Promoter Score) according to the embodiment. NPS is an index for quantifying customer loyalty to a predetermined product or service. NPS can also be rephrased as an index for quantifying customer satisfaction with a predetermined product or service. For example, a questionnaire is conducted for customers who have purchased a product, asking "To what extent would you like to recommend this product?" and having the customers select a score from 0 to 10 indicating the degree to which they would like to recommend the product to others and answer. Then, the customers select and answer a score corresponding to the degree to which they would like to recommend the product to others. Hereinafter, in the questionnaire responses from the customers, the score selected by the customers may be referred to as "NPS data".

[0011] Here, customers who selected a score of 0 to 6 are presumed to be customers with a low likelihood of recommending the product to others, so they are regarded as critics of the product. Also, customers who selected a score of 7 to 8 are presumed to be customers with a medium likelihood of recommending the product to others, so they are regarded as neutrals regarding the product. Further, customers who selected a score of 9 to 10 are presumed to be customers with a high likelihood of recommending the product to others, so they are regarded as recommenders of the product. Also, NPS is calculated by subtracting the percentage of customers who are critics from the percentage of customers who are recommenders.

[0012] Figure 2 is a diagram for explaining the problems of variable importance related to the comparative technique. In the comparative technique, based on a plurality of set data that are pairs of customer data regarding customers and the NPS data of those customers, a decision tree model is generated that is learned to output the NPS data of the customer corresponding to the target variable when the customer data corresponding to the explanatory variable is input. Here, the customer data includes attribute information regarding the attributes of the customer (hereinafter, may be referred to as "customer attribute information"). For example, the customer attribute information may include information indicating demographic attributes such as the age group of the customer (teenagers, twenties, etc. Note that it may also be age), gender (male, female, etc. Hereinafter, it shall be male or female), occupation, place of residence, etc. Also, the customer data includes non-attribute information (hereinafter, may be referred to as "customer non-attribute information") which is information regarding the customer and is of a different type from the customer's attributes. For example, the customer non-attribute information may include information indicating whether the customer holds a credit card of a specific company or information indicating whether the customer subscribes to a specific service, etc. In the comparative technique, a decision tree model is generated that is learned to output the NPS data of the customer corresponding to the target variable when customer data including both the customer's attribute information and non-attribute information is input at once. And the variable importance generated when the decision tree model is generated is obtained.

[0013] FIG. 2 shows a state where the variable importance of explanatory variables corresponding to attribute information regarding attributes such as the age and gender of customers and the variable importance of explanatory variables corresponding to non-attribute information regarding non-attributes different from attributes such as whether a customer subscribes to the store smartphone support service and whether the customer subscribes to the credit card of Company A are output at once. Further, in FIG. 2, as a problem with the variable importance related to the comparative technique, for example, it is considered that there is a correlation between an explanatory variable corresponding to attribute information such as the age of a customer and an explanatory variable corresponding to non-attribute information such as whether the customer subscribes to the credit card of Company A. For example, it is considered that there is a correlation that the older the customer is, the higher the possibility of subscribing to the credit card of Company A. Thus, in the comparative technique, since the dataset used to train the decision tree model contains a correlation (also called multicollinearity) of explanatory variables, it may be difficult to accurately estimate important explanatory variables for the target variable.

[0014] FIG. 3 is a diagram for explaining an outline of information processing according to an embodiment. In FIG. 3, an information processing apparatus according to the embodiment acquires a plurality of set data that are pairs of explanatory data that are customer data and target data that are NPS data of customers. Further, the information processing apparatus receives from a user a designation of an explanatory variable corresponding to the attribute information of a customer and an explanatory variable corresponding to the non-attribute information of the customer. Further, when the attribute information of a customer corresponding to each of the explanatory variables received as the explanatory variables corresponding to the attribute information of a customer is input, the information processing apparatus generates a first decision tree model M1 that is learned to output the NPS data of the customer. Further, the information processing apparatus, among the customer data included in each of the set data corresponding to each of the leaf nodes of the first decision tree model M1, when the non-attribute information of a customer corresponding to each of the explanatory variables received as the explanatory variables corresponding to the non-attribute information of the customer is input, generates second decision tree models M2-1 to M2-N (N is a natural number) that are learned to output the NPS data of the customers included in each of the set data corresponding to each of the leaf nodes of the first decision tree model M1. When it is not necessary to distinguish the second decision tree models M2-1 to M2-N, they are described as the second decision tree model M2.

[0015] As described with reference to FIG. 3, the information processing apparatus according to the embodiment performs learning in two stages: learning corresponding to explanatory variables corresponding to customer attribute information and learning corresponding to explanatory variables corresponding to customer non-attribute information. Specifically, the information processing apparatus generates a first decision tree model that has learned only the degree of contribution (e.g., importance) of the explanatory variables corresponding to the customer attribute information to the objective variable by learning the first decision tree model using only the explanatory data corresponding to the explanatory variables corresponding to the customer attribute information. Subsequently, the information processing apparatus learns a second decision tree model corresponding to explanatory variables corresponding to customer non-attribute information different from the explanatory variables corresponding to customer attribute information, based on the set data classified into each of the leaf nodes of the first decision tree model that has learned only the degree of contribution of the explanatory variables corresponding to the customer attribute information to the objective variable. That is, the information processing apparatus generates a second decision tree model that has learned the degree of contribution (e.g., importance) of the explanatory variables corresponding to the customer non-attribute information to the objective variable, after eliminating the influence of the contribution of the explanatory variables corresponding to the customer attribute information to the objective variable.

[0016] Thereby, the information processing apparatus can calculate the variable importance indicating the degree of contribution of the explanatory variables corresponding to the customer non-attribute information to the objective variable, without being affected by the explanatory variables corresponding to the customer attribute information. That is, the information processing apparatus can calculate the variable importance of the explanatory variables corresponding to the customer non-attribute information, in a state where the influence of the correlation between the explanatory variables corresponding to the customer attribute information and the explanatory variables corresponding to the customer non-attribute information is eliminated. Thereby, the information processing apparatus can accurately estimate important explanatory variables for the objective variable from among the explanatory variables corresponding to the customer non-attribute information, based on the variable importance of the explanatory variables corresponding to the customer non-attribute information, in a state where the influence of the explanatory variables corresponding to the customer attribute information is eliminated. Therefore, the information processing apparatus can accurately estimate important explanatory variables for the objective variable.

[0017] Note that, hereinafter, the customer is referred to as the "target person to be analyzed". Also, hereinafter, the customer data is referred to as the "information regarding the target person to be analyzed". In addition, in the present embodiment, the case where the target data corresponding to the target variable is NPS data will be described, but the target data corresponding to the target variable is not limited to NPS data. For example, the target data corresponding to the target variable may be any data as long as it is data of the target that the user desires to analyze.

[0018] [2. Configuration of Information Processing Apparatus] FIG. 4 is a diagram showing a configuration example of an information processing apparatus according to an embodiment. The information processing apparatus 100 according to the embodiment includes a communication unit 110, a storage unit 120, and a control unit 130.

[0019] (Communication Unit 110) The communication unit 110 is realized by, for example, a NIC (Network Interface Card), an antenna, or the like. Specifically, the communication unit 110 is connected to a network N (not shown) by wire or wirelessly, and transmits and receives information to and from a terminal device used by a user.

[0020] (Storage Unit 120) The storage unit 120 is realized by, for example, a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disk. Specifically, the storage unit 120 stores an information processing program according to the embodiment.

[0021] (Control Unit 130) The control unit 130 is a controller, which is realized, for example, by a CPU (Central Processing Unit), an MPU (Micro Processing Unit), etc., when various programs (for example, the information processing program according to the embodiment) stored in the storage device inside the information processing apparatus 100 are executed with the RAM as a working area. Further, the control unit 130 is a controller and is realized, for example, by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

[0022] The control unit 130 has a reception unit 131, a generation unit 132, an acquisition unit 133, and a provision unit 134 as functional units, and may realize or execute the operations of information processing described below. Note that the internal configuration of the control unit 130 is not limited to the configuration shown in FIG. 4, and any other configuration may be used as long as it can perform the information processing described later. Further, each functional unit indicates the function of the control unit 130 and does not necessarily have to be physically distinct.

[0023] (Reception Unit 131) The reception unit 131 receives various types of information. Specifically, the reception unit 131 receives a designation of a first type of explanatory variable and a second type of explanatory variable different from the first type of explanatory variable from the user. For example, the reception unit 131 displays on the screen of the terminal device used by the user the first type of explanatory variable in a selectable state from among a plurality of types of explanatory variables. For example, as an example of the first type of explanatory variable, the reception unit 131 displays in a selectable state an explanatory variable with a low priority for obtaining the variable importance by the user. Also, for example, as an example of the first type of explanatory variable, the reception unit 131 displays in a selectable state an explanatory variable corresponding to the attribute information (hereinafter, may be referred to as "attribute information of the target person") regarding the attributes of the target person to be analyzed. For example, the attribute information of the target person may include information indicating demographic attributes such as the age group of the target person (teenagers, twenties, etc. Note that it may also be the age), gender (male, female, etc. Hereinafter, it is assumed to be male or female), occupation, place of residence, etc.

[0024] Subsequently, the reception unit 131 receives from the terminal device information regarding the explanatory variable selected as the first type of explanatory variable by the user from among a plurality of types of explanatory variables. By receiving the information regarding the explanatory variable selected as the first type of explanatory variable by the user, the reception unit 131 receives the designation of the first type of explanatory variable from the user. For example, the first type of explanatory variable is an explanatory variable corresponding to the attribute information regarding the attributes of the target person to be analyzed. Alternatively, the first type of explanatory variable may be an explanatory variable designated as an explanatory variable with a low priority for obtaining the variable importance by the user.

[0025] Note that the reception unit 131 may receive information regarding explanatory variables not selected as the first type of explanatory variables by the user from the terminal device among a plurality of types of explanatory variables. The reception unit 131 may receive information regarding explanatory variables not selected as the first type of explanatory variables by the user as information regarding the second type of explanatory variables. That is, the reception unit 131 may receive a designation of the second type of explanatory variables from the user by receiving information regarding explanatory variables not selected as the first type of explanatory variables by the user.

[0026] In addition, the reception unit 131 displays, on the screen of the terminal device used by the user, the second type of explanatory variables selectable from among a plurality of types of explanatory variables. For example, the reception unit 131 displays, as an example of the second type of explanatory variables, explanatory variables with a high priority for obtaining variable importance by the user in a selectable state. Also, for example, the reception unit 131 displays, as an example of the second type of explanatory variables, explanatory variables corresponding to non-attribute information (hereinafter, may be referred to as "non-attribute information of the target person") that is information regarding the target person to be analyzed and is of a type different from the attributes of the target person. For example, the non-attribute information of the target person may include information indicating non-attributes other than information indicating attributes such as the age, gender, occupation, place of residence, etc. of the target person. For example, the non-attribute information of the target person may include, as information indicating non-attributes, information indicating whether the target person has a credit card of a specific company or information indicating whether the target person subscribes to a specific service.

[0027] Subsequently, the reception unit 131 receives, from the terminal device, information regarding the explanatory variable selected by the user as the second type of explanatory variable from among a plurality of types of explanatory variables. By receiving the information regarding the explanatory variable selected by the user as the second type of explanatory variable, the reception unit 131 accepts the designation of the second type of explanatory variable from the user. For example, the second type of explanatory variable is an explanatory variable corresponding to non-attribute information, which is information regarding the subject to be analyzed and is of a type different from the attributes of the subject. Alternatively, the second type of explanatory variable may be an explanatory variable designated by the user as an explanatory variable having a high priority for obtaining variable importance.

[0028] Note that the reception unit 131 may receive, from the terminal device, information regarding the explanatory variables not selected by the user as the second type of explanatory variables among the plurality of types of explanatory variables. The reception unit 131 may accept the information regarding the explanatory variables not selected by the user as the second type of explanatory variables as the information regarding the first type of explanatory variables. That is, the reception unit 131 may accept the designation of the first type of explanatory variables from the user by receiving the information regarding the explanatory variables not selected by the user as the second type of explanatory variables.

[0029] (Generation unit 132) The generation unit 132 generates various types of information. Specifically, the generation unit 132 generates, based on a plurality of sets of data, which are pairs of explanatory data corresponding to each of the plurality of types of explanatory variables and target data corresponding to the target variable, a first decision tree model that predicts the target data from the first explanatory data corresponding to each of the first type of explanatory variables among the plurality of types of explanatory variables, and generates, based on the set of data corresponding to each of the leaf nodes of the generated first decision tree model, a second decision tree model that predicts the target data from the second explanatory data corresponding to the second type of explanatory variables, which is different from the first type of explanatory variables, among the plurality of types of explanatory variables.

[0030] Here, the first decision tree model or the second decision tree model according to the present embodiment is a machine learning model generated using a known decision tree algorithm such as CART, CHAID, C5.0, etc. For example, the first decision tree model or the second decision tree model may be a classification tree or a regression tree. Further, the first decision tree model or the second decision tree model may be a machine learning model generated by an ensemble method based on a decision tree algorithm. For example, the first decision tree model or the second decision tree model may be a machine learning model generated by an ensemble method that combines a decision tree algorithm with random forest or a decision tree algorithm with gradient boosting.

[0031] More specifically, the storage unit 120 stores a plurality of set data that are sets of explanatory data corresponding to each of a plurality of types of explanatory variables and target data corresponding to the target variable. For example, the storage unit 120 stores a plurality of set data that are sets of explanatory data that is subject data regarding the subject to be analyzed and target data that is NPS data (hereinafter, may be referred to as "NPS data of the subject") that is an answer to a questionnaire for the subject. Note that the subject data includes both the attribute information of the subject and the non-attribute information of the subject. Further, the generation unit 132 acquires a plurality of set data that are sets of explanatory data corresponding to each of a plurality of types of explanatory variables and target data corresponding to the target variable. For example, the generation unit 132 acquires a plurality of set data that are sets of explanatory data that is subject data and target data that is NPS data of the subject. For example, the generation unit 132 acquires a plurality of set data by referring to the storage unit 120.

[0032] Subsequently, when the generation unit 132 acquires a plurality of sets of data, based on the plurality of sets of data, among the plurality of types of explanatory variables, it generates a first decision tree model that predicts target data from the first explanatory data corresponding to each of the first type of explanatory variables. Specifically, the generation unit 132 generates a first decision tree model that is learned to predict the target data included in each of the plurality of sets of data from the first explanatory data corresponding to each of the explanatory variables received by the reception unit 131 as the first type of explanatory variables among the explanatory data included in each of the plurality of sets of data. For example, the generation unit 132 generates a first decision tree model that is learned to predict the NPS data of the target person from the attribute information of the target person received by the reception unit 131 as the first type of explanatory variables among the target person data included in each of the plurality of sets of data.

[0033] For example, when the first explanatory data corresponding to each of the explanatory variables received by the reception unit 131 as the first type of explanatory variables is input among the explanatory data included in each of the plurality of sets of data, the generation unit 132 generates a first decision tree model that is learned to output the target data included in each of the plurality of sets of data. For example, when the attribute information of the target person is input among the target person data included in each of the plurality of sets of data, the generation unit 132 generates a first decision tree model that is learned to output the NPS data of the target person included in each of the plurality of sets of data. Further, when the generation unit 132 generates the first decision tree model, it associates information that can identify the first decision tree model with information regarding the first decision tree model and stores the result in the storage unit 120.

[0034] For example, when a group of attribute information of a target person is input, the generation unit 132 generates a first decision tree including a root node corresponding to the group of attribute information of the target person. For example, when a group of attribute information of a target person is input, the generation unit 132 generates a first branching condition for dividing the group of attribute information of the target person into two groups based on whether the gender of the target person is male or female. For example, when the generation unit 132 generates the first branching condition, the generation unit 132 divides the group of attribute information of the target person into two groups: a group of attribute information of a target person whose gender is male (hereinafter, may be referred to as "a group of attribute information of a male target person"), and a group of attribute information of a target person whose gender is female (hereinafter, may be referred to as "a group of attribute information of a female target person"). Further, the generation unit 132 generates a first decision tree including a male node corresponding to the group of attribute information of a male target person, an edge branching from the root node to the male node, a female node corresponding to the group of attribute information of a female target person, and an edge branching from the root node to the female node.

[0035] Subsequently, when the generation unit 132 divides the group of the target person's attribute information into two groups, i.e., the group of the male target person's attribute information and the group of the female target person's attribute information, for example, based on whether the age of the target person is 50 or older, the generation unit 132 generates a second condition for dividing the group of the male target person's attribute information into two groups. For example, when the generation unit 132 generates the second condition, the group of the male target person's attribute information is divided into a group of the male target person's attribute information whose age is 50 or older (hereinafter, may be described as "the group of the male target person's attribute information aged 50 or older"), and a group of the male target person's attribute information whose age is less than 50 (hereinafter, may be described as "the group of the male target person's attribute information aged less than 50"). Further, the generation unit 132 generates a first decision tree including a male node aged 50 or older corresponding to the group of the male target person's attribute information aged 50 or older, an edge branching from the male node to the male node aged 50 or older, a male node aged less than 50 corresponding to the group of the male target person's attribute information aged less than 50, and an edge branching from the male node to the male node aged less than 50.

[0036] In addition, when the generation unit 132 divides the group of attribute information of the target person into two groups, namely, the group of attribute information of male target persons and the group of attribute information of female target persons, for example, based on whether the age of the target person is 40 or older, the generation unit 132 generates a third condition for dividing the group of attribute information of female target persons into two groups. For example, when the generation unit 132 generates the third condition, the group of attribute information of female target persons is divided into a group of attribute information of female target persons whose age is 40 or older (hereinafter, may be described as "the group of attribute information of female target persons aged 40 or older"), and a group of attribute information of female target persons whose age is less than 40 (hereinafter, may be described as "the group of attribute information of female target persons under 40"). Further, the generation unit 132 generates a first decision tree including a node for women aged 40 or older corresponding to the group of attribute information of female target persons aged 40 or older, an edge branching from the women node to the node for women aged 40 or older, a node for women under 40 corresponding to the group of attribute information of female target persons under 40, and an edge branching from the women node to the node for women under 40.

[0037] Here, the generation unit 132 generating a predetermined branching condition corresponds to the degree of contribution of the explanatory variable corresponding to the predetermined branching condition to the objective variable being high. In other words, the generation unit 132 generating a predetermined branching condition means that the explanatory data corresponding to the explanatory variable corresponding to the predetermined branching condition is important in predicting the objective data corresponding to the objective variable. That is, the generation unit 132 generating a predetermined branching condition corresponds to the high importance of the explanatory variable corresponding to the predetermined branching condition. Also, the high importance of the explanatory variable corresponds to the high variable importance corresponding to the explanatory variable. Here, the variable importance is an index representing the degree of contribution of the explanatory variable to the objective variable. In other words, the variable importance is an index representing how much the explanatory data corresponding to the explanatory variable has influenced in predicting the objective data corresponding to the objective variable. That is, a high variable importance means that the explanatory data corresponding to the explanatory variable has a large influence in predicting the objective data corresponding to the objective variable.

[0038] For example, the generation unit 132 generating a first branching condition regarding whether the target person is male or female means that the importance of the explanatory variable corresponding to the gender of the target person is high. Also, the high importance of the explanatory variable corresponding to the gender of the target person means that the explanatory data corresponding to the gender of the target person has a great influence in predicting the NPS data of the target person.

[0039] Also, the generation unit 132 generating a second branching condition regarding whether the age of the male target person is 50 or older means that the importance of the explanatory variable corresponding to the age of the male target person is high. Also, the high importance of the explanatory variable corresponding to the age of the male target person means that the explanatory data corresponding to the age of the male target person has a great influence in predicting the NPS data of the male target person.

[0040] Also, the generation unit 132 generating a third branching condition regarding whether the age of the female target person is 40 or older means that the importance of the explanatory variable corresponding to the age of the female target person is high. Also, the high importance of the explanatory variable corresponding to the age of the female target person means that the explanatory data corresponding to the age of the female target person has a great influence in predicting the NPS data of the female target person.

[0041] In addition, when generating the first decision tree model, the generation unit 132 generates a first variable importance. For example, when the branch condition of the decision tree is generated, the generation unit 132 may calculate the first variable importance corresponding to each of the first type of explanatory variables by counting the number of times the explanatory variable is used. Further, the generation unit 132 may calculate the first variable importance corresponding to each of the first type of explanatory variables based on the amount of decrease in Gini impurity in the explanatory data corresponding to each node of the decision tree. In this way, when generating the first decision tree model, the generation unit 132 calculates the first variable importance corresponding to each of the first type of explanatory variables, thereby generating the first variable importance corresponding to each of the first type of explanatory variables. Further, when the generation unit 132 generates the first variable importance corresponding to each of the first type of explanatory variables, the generation unit 132 associates the information that can identify the first decision tree model, the information that can identify each of the first type of explanatory variables, and the first variable importance corresponding to each of the first type of explanatory variables, and stores the associated information in the storage unit 120.

[0042] Subsequently, when the generation unit 132 generates the first decision tree model, based on the set data corresponding to each leaf node of the generated first decision tree model, among the plurality of types of explanatory variables, a second decision tree model for predicting target data from the second explanatory data corresponding to the second type of explanatory variables different from the first type of explanatory variables is generated. Here, the set data corresponding to each leaf node of the first decision tree model is, in other words, the set data divided by each of the branch conditions corresponding to each node from the root node to the leaf node of the first decision tree model.

[0043] Specifically, the generation unit 132 generates a second decision tree model that is learned to predict the target data included in each set of data corresponding to each leaf node of the first decision tree model from the second explanatory data corresponding to each explanatory variable received as the second type of explanatory variable by the reception unit 131. For example, when the second explanatory data corresponding to each explanatory variable received as the second type of explanatory variable by the reception unit 131 is input among the explanatory data included in each set of data corresponding to each leaf node of the first decision tree model, the generation unit 132 generates a second decision tree model that is learned to output the target data included in each set of data corresponding to each leaf node of the first decision tree model. For example, when the second explanatory data corresponding to the non-attribute information of the subject is input among the explanatory data that is subject data, the generation unit 132 generates a second decision tree model that is learned to output the target data that is the NPS data of the subject. In this way, the generation unit 132 generates a second decision tree model corresponding to each leaf node of the first decision tree model based on each set of data corresponding to each leaf node of the first decision tree model. Further, when the generation unit 132 generates the second decision tree model, the generation unit 132 associates information that can identify the second decision tree model, information regarding the second decision tree model, and information that can identify the first decision tree model that is the basis for generating the second decision tree model, and stores the associated information in the storage unit 120.

[0044] For example, assume that the leaf nodes of the above-described first decision tree model correspond to a node for men aged 50 or older corresponding to a group of attribute information of subjects who are men aged 50 or older, a node for men under 50 corresponding to a group of attribute information of subjects who are men under 50, a node for women aged 40 or older corresponding to a group of attribute information of subjects who are women aged 40 or older, and a node for women under 40 corresponding to a group of attribute information of subjects who are women under 40. At this time, the generation unit 132 generates each of the four second decision tree models M21 to M24 based on the set data corresponding to each of the node for men aged 50 or older, the node for men under 50, the node for women aged 40 or older, and the node for women under 40.

[0045] For example, the generation unit 132 generates the second decision tree model M21 based on the set data corresponding to the group of attribute information of subjects who are men aged 50 or older. For example, among the subject data included in each of the set data corresponding to the group of attribute information of subjects who are men aged 50 or older, the generation unit 132 generates the second decision tree model M21 that is learned to predict the NPS data of the subject from the non-attribute information of the subject received by the reception unit 131 as the second type of explanatory variable.

[0046] In addition, the generation unit 132 generates the second decision tree model M22 based on the set data corresponding to the group of attribute information of subjects who are men under 50. For example, among the subject data included in each of the set data corresponding to the group of attribute information of subjects who are men under 50, the generation unit 132 generates the second decision tree model M22 that is learned to predict the NPS data of the subject from the non-attribute information of the subject received by the reception unit 131 as the second type of explanatory variable.

[0047] Further, the generation unit 132 generates a second decision tree model M23 based on the set data corresponding to the group of attribute information of the target persons who are women aged 40 or older. For example, the generation unit 132 predicts the NPS data of the target persons from the non-attribute information of the target persons received as the second type of explanatory variable by the reception unit 131 among the target person data included in each of the set data corresponding to the group of attribute information of the target persons who are women aged 40 or older, and generates a second decision tree model M23 that has been learned to do so.

[0048] Also, the generation unit 132 generates a second decision tree model M24 based on the set data corresponding to the group of attribute information of the target persons who are women under 40 years old. For example, the generation unit 132 predicts the NPS data of the target persons from the non-attribute information of the target persons received as the second type of explanatory variable by the reception unit 131 among the target person data included in each of the set data corresponding to the group of attribute information of the target persons who are women under 40 years old, and generates a second decision tree model M24 that has been learned to do so.

[0049] Further, when generating the second decision tree model, the generation unit 132 generates a second variable importance. For example, when the branching condition of the decision tree is generated, the generation unit 132 may calculate the second variable importance corresponding to each of the second type of explanatory variables by counting the number of times the explanatory variable is used. Also, the generation unit 132 may calculate the second variable importance corresponding to each of the second type of explanatory variables based on the decrease in Gini impurity in the explanatory data corresponding to each node of the decision tree. In this way, the generation unit 132 generates the second variable importance by calculating the second variable importance corresponding to each of the second type of explanatory variables when generating the second decision tree model. Also, when the generation unit 132 generates the second variable importance corresponding to each of the second type of explanatory variables, it associates the information that can identify the second decision tree model, the information that can identify each of the second type of explanatory variables, and the second variable importance corresponding to each of the second type of explanatory variables, and stores them in the storage unit 120.

[0050] (Acquisition unit 133) The acquisition unit 133 acquires various types of information. Specifically, the acquisition unit 133 acquires the first variable importance that is generated when the generation unit 132 generates the first decision tree model. For example, the acquisition unit 133 refers to the storage unit 120 and acquires the first variable importance corresponding to each of the first type of explanatory variables associated with information that can identify the first decision tree model when the generation unit 132 generates the first decision tree model. Further, the acquisition unit 133 acquires the second variable importance that is generated when the generation unit 132 generates the second decision tree model. For example, the acquisition unit 133 refers to the storage unit 120 and acquires the second variable importance corresponding to each of the second type of explanatory variables associated with information that can identify the second decision tree model when the generation unit 132 generates the second decision tree model.

[0051] FIG. 5 is a diagram for explaining an example of information processing according to an embodiment. FIG. 5 illustrates information processing E1 according to the embodiment. In FIG. 5, an acquisition unit 133 acquires a first predicted value by a first decision tree model M1 generated by a generation unit 132. For example, the acquisition unit 133 refers to a storage unit 120 to acquire information regarding the first decision tree model M1 generated by the generation unit 132. Further, the acquisition unit 133 refers to the storage unit 120 to acquire verification set data. Here, the verification set data is a set of verification explanatory data corresponding to each of a plurality of types of explanatory variables and verification target data corresponding to a target variable. Also, the verification set data is the same type of set data as a plurality of sets of data (hereinafter, may be referred to as "learning set data") used when generating the first decision tree model and the second decision tree model. When the acquisition unit 133 acquires information regarding the first decision tree model M1 and the verification set data, the acquisition unit 133 inputs, into the first decision tree model M1, first explanatory data corresponding to each of the explanatory variables received as the first type of explanatory variable by a reception unit 131 among the explanatory data included in each of the verification set data, thereby acquiring a first predicted value. Subsequently, when the acquisition unit 133 acquires the first predicted value, the acquisition unit 133 calculates a first error, which is the error between the verification target data included in the verification set data and the first predicted value, and verifies the prediction accuracy of the first decision tree model M1 based on the calculated first error.

[0052] Also, in FIG. 5, the acquisition unit 133 acquires a first variable importance corresponding to the first decision tree model M1 whose first error is less than a first threshold value. For example, the acquisition unit 133 calculates a first error corresponding to each of a plurality of different first decision tree models M1 generated by the generation unit 132. Subsequently, the acquisition unit 133 selects, from among the plurality of different first decision tree models M1, a first decision tree model M1 whose first error is less than the first threshold value, and acquires the first variable importance corresponding to the selected first decision tree model M1.

[0053] Further, the acquisition unit 133 acquires the second predicted value by the second decision tree model generated by the generation unit 132. In FIG. 5, the generation unit 132 generates second decision tree models M2-1 to M2-3 corresponding to each of the leaf nodes of the first decision tree model M1 based on the set data corresponding to each of the leaf nodes of the first decision tree model M1. For example, the acquisition unit 133 refers to the storage unit 120 and acquires information on the second decision tree models M2-1 to M2-3 generated by the generation unit 132. Further, the acquisition unit 133 acquires verification set data corresponding to each of the leaf nodes of the first decision tree model M1. For example, the acquisition unit 133 acquires verification set data corresponding to the leaf nodes corresponding to each of the second decision tree models M2-1 to M2-3. For example, when the acquisition unit 133 acquires information on the second decision tree model M2-1 and verification set data corresponding to the leaf node corresponding to the second decision tree model M2-1, among the explanatory data included in each of the verification set data, the second explanatory data corresponding to each of the explanatory variables received as the second type of explanatory variable by the reception unit 131 is input to the second decision tree model M2-1, thereby acquiring the second predicted value by the second decision tree model M2-1. Further, when the acquisition unit 133 acquires the second predicted value by the second decision tree model M2-1, the acquisition unit 133 calculates a second error by the second decision tree model M2-1, which is the error between the verification target data included in the verification set data and the second predicted value by the second decision tree model M2-1, and verifies the prediction accuracy of the second decision tree model M2-1 based on the calculated second error by the second decision tree model M2-1.

[0054] In addition, when the acquisition unit 133 acquires information on the second decision tree model M2-2 and verification set data corresponding to the leaf node corresponding to the second decision tree model M2-2, among the explanatory data included in each of the verification set data, the second explanatory data corresponding to each of the explanatory variables accepted as the second type of explanatory variable by the reception unit 131 is input to the second decision tree model M2-2, thereby obtaining a second predicted value by the second decision tree model M2-2. Further, when the acquisition unit 133 obtains the second predicted value by the second decision tree model M2-2, it calculates a second error by the second decision tree model M2-2, which is the error between the verification target data included in the verification set data and the second predicted value by the second decision tree model M2-2, and verifies the prediction accuracy of the second decision tree model M2-2 based on the calculated second error by the second decision tree model M2-2.

[0055] In addition, when the acquisition unit 133 acquires information on the second decision tree model M2-3 and verification set data corresponding to the leaf node corresponding to the second decision tree model M2-3, among the explanatory data included in each of the verification set data, the second explanatory data corresponding to each of the explanatory variables accepted as the second type of explanatory variable by the reception unit 131 is input to the second decision tree model M2-3, thereby obtaining a second predicted value by the second decision tree model M2-3. Further, when the acquisition unit 133 obtains the second predicted value by the second decision tree model M2-3, it calculates a second error by the second decision tree model M2-3, which is the error between the verification target data included in the verification set data and the second predicted value by the second decision tree model M2-3, and verifies the prediction accuracy of the second decision tree model M2-3 based on the calculated second error by the second decision tree model M2-3.

[0056] Also, in FIG. 5, the acquisition unit 133 acquires the second variable importance corresponding to the second decision tree model in which the second error is less than the second threshold. Specifically, the acquisition unit 133 acquires the second variable importance corresponding to each of the second decision tree models in which each of the second errors by the second decision tree models corresponding to the leaf nodes of the first decision tree model generated by the generation unit 132 is less than the second threshold. For example, the acquisition unit 133 calculates the second error corresponding to each of a plurality of different second decision tree models M2-1 generated by the generation unit 132. Subsequently, the acquisition unit 133 selects a second decision tree model M2-1 in which the second error is less than the second threshold from among the plurality of different second decision tree models M2-1, and acquires the second variable importance corresponding to the selected second decision tree model M2-1. Also, the acquisition unit 133 calculates the second error corresponding to each of a plurality of different second decision tree models M2-2 generated by the generation unit 132. Subsequently, the acquisition unit 133 selects a second decision tree model M2-2 in which the second error is less than the second threshold from among the plurality of different second decision tree models M2-2, and acquires the second variable importance corresponding to the selected second decision tree model M2-2. Also, the acquisition unit 133 calculates the second error corresponding to each of a plurality of different second decision tree models M2-3 generated by the generation unit 132. Subsequently, the acquisition unit 133 selects a second decision tree model M2-3 in which the second error is less than the second threshold from among the plurality of different second decision tree models M2-3, and acquires the second variable importance corresponding to the selected second decision tree model M2-3.

[0057] Further, the acquisition unit 133 acquires the second variable importance corresponding to each of the second decision tree models generated by the generation unit 132, and calculates the total importance, which is the variable importance corresponding to each of the second types of explanatory variables, by summing up the acquired second variable importance for each of the second types of explanatory variables. In FIG. 5, the acquisition unit 133 acquires the second variable importance corresponding to each of the second decision tree models M2-1 to M2-3 generated by the generation unit 132, and calculates the total importance, which is the variable importance corresponding to each of the second types of explanatory variables, by summing up the acquired second variable importance for each of the second types of explanatory variables. For example, the acquisition unit 133 calculates the total importance, which is the variable importance corresponding to each of the second types of explanatory variables, by summing up the second variable importance corresponding to the second decision tree model M2-1, the second variable importance corresponding to the second decision tree model M2-2, and the second variable importance corresponding to the second decision tree model M2-3 for each of the second types of explanatory variables.

[0058] Further, the acquisition unit 133 acquires the second predicted values corresponding to each of the second decision tree models generated by the generation unit 132, and calculates the combined predicted value, which is the overall predicted value of the second decision tree models generated by the generation unit 132, by combining the acquired second predicted values. For example, the acquisition unit 133 calculates the combined predicted value by calculating the average of the second predicted values corresponding to each of the second decision tree models generated by the generation unit 132.

[0059] FIG. 6 is a diagram for explaining an example of information processing according to an embodiment. The generation unit 132 generates a plurality of different first decision tree models by varying parameters when generating a decision tree model, and generates a plurality of different second decision tree models corresponding to each of the plurality of generated different first decision tree models. In FIG. 6, as a parameter when generating a decision tree model, by varying the seed value when generating a decision tree model each time, a state in which the information processing shown in FIG. 5 is repeated N times (N is a natural number of 2 or more) is shown. Specifically, the generation unit 132 generates N different first decision tree models by varying the seed value when generating the first decision tree model each time. Further, the generation unit 132 generates a plurality of different second decision tree models corresponding to each of the N different first decision tree models generated by varying the seed value when generating the second decision tree model each time.

[0060] Further, the acquisition unit 133 calculates the average of the first variable importance corresponding to each of the plurality of different first decision tree models generated by the generation unit 132 for each first type of explanatory variable. In FIG. 6, the acquisition unit 133 calculates the average of the first variable importance corresponding to each of the N different first decision tree models generated by the generation unit 132 for each first type of explanatory variable.

[0061] Further, the acquisition unit 133 calculates the average of the total importance corresponding to each of the plurality of different second decision tree models generated by the generation unit 132 for each second type of explanatory variable. In FIG. 6, the acquisition unit 133 calculates the average of the total importance corresponding to each of the N different second decision tree models generated by the generation unit 132 for each second type of explanatory variable.

[0062] Further, the acquisition unit 133 calculates the average of the first predicted values corresponding to each of the plurality of different first decision tree models generated by the generation unit 132. In FIG. 6, the acquisition unit 133 calculates the average of the first predicted values corresponding to each of the N different first decision tree models generated by the generation unit 132.

[0063] Further, the acquisition unit 133 calculates the average of the combined prediction values corresponding to each of the plurality of different second decision tree models generated by the generation unit 132. In FIG. 6, the acquisition unit 133 calculates the average of the combined prediction values corresponding to each of the N different second decision tree models generated by the generation unit 132.

[0064] (Providing unit 134) The providing unit 134 provides various information. Specifically, the providing unit 134 provides the user with information regarding the average of the first variable importance, the average of the total importance, the average of the first prediction values, and the average of the combined prediction values calculated by the acquisition unit 133. For example, the providing unit 134 causes a terminal device to display information regarding the average of the first variable importance, the average of the total importance, the average of the first prediction values, and the average of the combined prediction values on the screen.

[0065] [3. Processing procedure] FIG. 7 is a flowchart showing a processing procedure by the information processing apparatus according to the embodiment. In FIG. 7, the generation unit 132 of the information processing apparatus 100 generates a plurality of different first decision tree models by varying the parameters when generating the decision tree model (step S101). Subsequently, the generation unit 132 of the information processing apparatus 100 generates a plurality of different second decision tree models corresponding to each of the plurality of different first decision tree models generated by varying the parameters when generating the decision tree model (step S102).

[0066] Further, the acquisition unit 133 of the information processing apparatus 100 acquires the first variable importance corresponding to each of the plurality of different first decision tree models and the second variable importance corresponding to each of the plurality of different second decision tree models (step S103). For example, the acquisition unit 133 of the information processing apparatus 100 acquires the total importance as the second variable importance corresponding to each of the plurality of different second decision tree models. Subsequently, the acquisition unit 133 of the information processing apparatus 100 calculates the average of the first variable importance and the average of the second variable importance (step S104). For example, the acquisition unit 133 of the information processing apparatus 100 calculates the average of the total importance as the average of the second variable importance.

[0067] 〔4. Modification Example〕 In the above-described embodiment, the case where the first type of explanatory variable corresponds to attribute information and the second type of explanatory variable corresponds to non-attribute information has been described. However, the combination of the first type of explanatory variable and the second type of explanatory variable is not limited to the combination of the explanatory variable corresponding to attribute information and the explanatory variable corresponding to non-attribute information. For example, when it is desired to predict the number of customers at an ice cream store, as explanatory variables that affect the number of customers, the temperature on the day of sale, the precipitation, the price of ice cream, the amount of advertising related to ice cream, etc. can be considered. At this time, for example, a user who wishes to predict the number of customers at an ice cream store desires to know the influence of explanatory variables that can be controlled manually, such as the price of ice cream and the amount of advertising related to ice cream, on the number of customers, rather than the influence of explanatory variables that cannot be controlled manually, such as the temperature and precipitation on the day of sale, on the number of customers. At this time, explanatory variables that cannot be controlled manually, such as the temperature and precipitation on the day of sale, are an example of explanatory variables with a low priority for obtaining variable importance by the user. Also, explanatory variables that can be controlled manually, such as the price of ice cream and the amount of advertising related to ice cream, are an example of explanatory variables with a low priority for obtaining variable importance by the user. That is, the first type of explanatory variable may be an explanatory variable that cannot be controlled manually, such as the temperature and precipitation on the day of sale. Also, the second type of explanatory variable may be an explanatory variable that can be controlled manually, such as the price of ice cream and the amount of advertising related to ice cream.

[0068] In the above-described embodiment, the reception unit 131 has been described as receiving from the user a designation of a first type of explanatory variable and a second type of explanatory variable different from the first type of explanatory variable. However, the process of receiving the first type of explanatory variable and the second type of explanatory variable is not limited to this. Specifically, the storage unit 120 stores variable information in which information indicating that it is the first type of explanatory variable and / or information indicating that it is the second type of explanatory variable are associated one-to-one with each of the plurality of types of explanatory variables. Instead of receiving from the user a designation of the first type of explanatory variable and the second type of explanatory variable, the reception unit 131 refers to the variable information in the storage unit 120 and receives, as the first type of explanatory variable, an explanatory variable associated one-to-one with the information indicating that it is the first type of explanatory variable, and receives, as the second type of explanatory variable, an explanatory variable associated one-to-one with the information indicating that it is the second type of explanatory variable.

[0069] 〔5. Effect〕 As described above, the information processing apparatus 100 according to the embodiment includes a generation unit 132 and an acquisition unit 133. The generation unit 132 generates, based on a plurality of sets of data that are pairs of explanatory data corresponding to each of the plurality of types of explanatory variables and target data corresponding to the target variable, a first decision tree model that predicts the target data from the first explanatory data corresponding to each of the first type of explanatory variables among the plurality of types of explanatory variables, and generates, based on the set of data corresponding to each leaf node of the generated first decision tree model, a second decision tree model that predicts the target data from the second explanatory data corresponding to a second type of explanatory variable different from the first type of explanatory variable among the plurality of types of explanatory variables. The acquisition unit 133 acquires a second variable importance generated when the generation unit 132 generates the second decision tree model.

[0070] In this way, the information processing apparatus 100 performs learning corresponding to the first type of explanatory variable and learning corresponding to the second type of explanatory variable in two stages. Specifically, the information processing apparatus 100 learns the first decision tree model only with the explanatory data corresponding to the first type of explanatory variable, thereby generating a first decision tree model that has learned only the degree of contribution (e.g., importance) of the first type of explanatory variable to the objective variable. Subsequently, the information processing apparatus 100 learns a second decision tree model corresponding to a second type of explanatory variable different from the first type of explanatory variable based on the set data classified into each of the leaf nodes of the first decision tree model that has learned only the degree of contribution of the first type of explanatory variable to the objective variable. That is, the information processing apparatus 100 generates a second decision tree model that has learned the degree of contribution (e.g., importance) of the second type of explanatory variable to the objective variable after excluding the influence of the contribution of the first type of explanatory variable to the objective variable. Thereby, the information processing apparatus 100 can calculate the variable importance indicating the degree of contribution of the second type of explanatory variable to the objective variable without being affected by the first type of explanatory variable. That is, the information processing apparatus 100 can calculate the variable importance of the second type of explanatory variable in a state where the influence of the correlation between the first type of explanatory variable and the second type of explanatory variable is excluded. Thereby, the information processing apparatus 100 can accurately estimate, for example, important explanatory variables for the objective variable from among the second type of explanatory variables based on the variable importance of the second type of explanatory variable in a state where the influence of the first type of explanatory variable is excluded. Therefore, the information processing apparatus 100 can accurately estimate important explanatory variables for the objective variable. Also, since the information processing apparatus 100 can accurately estimate important explanatory variables for the objective variable, it can contribute to the achievement of Goal 9, "Build the infrastructure for industry and innovation," of the Sustainable Development Goals (SDGs).

[0071] Further, the acquisition unit 133 acquires the second variable importance corresponding to each of the second decision tree models generated by the generation unit 132, and calculates the total importance, which is the variable importance corresponding to each of the second types of explanatory variables, by summing up the acquired second variable importance for each of the second types of explanatory variables.

[0072] Thereby, the information processing apparatus 100 can accurately estimate the explanatory variables important for the target variable from among the second types of explanatory variables based on the total importance, which is the variable importance corresponding to each of the second types of explanatory variables.

[0073] Also, the generation unit 132 generates a plurality of different first decision tree models by varying the parameters when generating the decision tree models, and generates a plurality of different second decision tree models corresponding to each of the plurality of different first decision tree models generated. The acquisition unit 133 calculates the average of the total importance corresponding to each of the plurality of different second decision tree models generated by the generation unit 132 for each of the second types of explanatory variables.

[0074] Thereby, the information processing apparatus 100 can accurately estimate the explanatory variables important for the target variable from among the second types of explanatory variables based on the average of the total importance calculated for each of the second types of explanatory variables.

[0075] Further, the acquisition unit 133 acquires the first predicted value by the first decision tree model generated by the generation unit 132, verifies the prediction accuracy of the first decision tree model based on the first error, which is the error between the verification target data included in the verification set data and the first predicted value, acquires the second predicted value by the second decision tree model generated by the generation unit 132, and verifies the prediction accuracy of the second decision tree model based on the second error, which is the error between the verification target data and the second predicted value.

[0076] As a result, the information processing apparatus 100 can acquire the variable importance corresponding to the decision tree model with high prediction accuracy, so that it is possible to accurately estimate important explanatory variables for the target variable from among the explanatory variables based on the more accurate variable importance.

[0077] In addition, the acquisition unit 133 acquires the second variable importance corresponding to the second decision tree model in which the second error is less than the second threshold value.

[0078] As a result, the information processing apparatus 100 acquires the variable importance corresponding to the decision tree model with high prediction accuracy, so that it is possible to accurately estimate important explanatory variables for the target variable from among the explanatory variables based on the more accurate variable importance.

[0079] In addition, the first type of explanatory variable is an explanatory variable corresponding to the attribute information regarding the attributes of the target person to be analyzed, and the second type of explanatory variable is information regarding the target person and is an explanatory variable corresponding to non-attribute information which is a different type of information from the attributes of the target person.

[0080] In this way, the information processing apparatus 100 performs learning in two stages: learning corresponding to explanatory variables corresponding to the attribute information of the target person and learning corresponding to explanatory variables corresponding to the non-attribute information of the target person. Specifically, the information processing apparatus 100 learns a first decision tree model only with explanatory data corresponding to explanatory variables corresponding to the attribute information of the target person, thereby generating a first decision tree model that has learned only the degree of contribution (e.g., importance) of the explanatory variables corresponding to the attribute information of the target person to the objective variable. Subsequently, the information processing apparatus 100 learns a second decision tree model corresponding to explanatory variables corresponding to non-attribute information of the target person that are different from the explanatory variables corresponding to the attribute information of the target person, based on the set data classified into each of the leaf nodes of the first decision tree model that has learned only the degree of contribution of the explanatory variables corresponding to the attribute information of the target person to the objective variable. That is, the information processing apparatus 100 generates a second decision tree model that has learned the degree of contribution (e.g., importance) of the explanatory variables corresponding to the non-attribute information of the target person to the objective variable, after excluding the influence of the contribution of the explanatory variables corresponding to the attribute information of the target person to the objective variable. Thereby, the information processing apparatus 100 can calculate the variable importance indicating the degree of contribution of the explanatory variables corresponding to the non-attribute information of the target person to the objective variable, without being affected by the explanatory variables corresponding to the attribute information of the target person. That is, the information processing apparatus 100 can calculate the variable importance of the explanatory variables corresponding to the non-attribute information of the target person, in a state where the influence of the correlation between the explanatory variables corresponding to the attribute information of the target person and the explanatory variables corresponding to the non-attribute information of the target person is excluded. Thereby, the information processing apparatus 100 can accurately estimate important explanatory variables for the objective variable from among the explanatory variables corresponding to the non-attribute information of the target person, based on the variable importance of the explanatory variables corresponding to the non-attribute information of the target person, in a state where the influence of the explanatory variables corresponding to the attribute information of the target person is excluded. Therefore, the information processing apparatus 100 can accurately estimate important explanatory variables for the objective variable.

[0081] The first type of explanatory variable is an explanatory variable designated as an explanatory variable with a low priority for the user to obtain variable importance, and the second type of explanatory variable is an explanatory variable designated as an explanatory variable with a high priority for the user to obtain variable importance.

[0082] In this way, the information processing apparatus 100 performs learning corresponding to an explanatory variable designated as an explanatory variable with a low priority for the user to obtain variable importance (hereinafter, may be described as a "low-priority explanatory variable") and learning corresponding to an explanatory variable designated as an explanatory variable with a high priority for the user to obtain variable importance (hereinafter, may be described as a "high-priority explanatory variable") in two stages. Specifically, the information processing apparatus 100 learns only the degree of contribution (for example, importance) of the low-priority explanatory variable to the target variable by learning the first decision tree model using only the explanatory data corresponding to the low-priority explanatory variable, and generates the first decision tree model that has learned only the degree of contribution of the low-priority explanatory variable to the target variable. Subsequently, the information processing apparatus 100 learns the second decision tree model corresponding to the high-priority explanatory variable based on the set data classified into each leaf node of the first decision tree model that has learned only the degree of contribution of the low-priority explanatory variable to the target variable. That is, the information processing apparatus 100 generates the second decision tree model that has learned the degree of contribution (for example, importance) of the high-priority explanatory variable to the target variable after excluding the influence of the contribution of the low-priority explanatory variable to the target variable. Thereby, the information processing apparatus 100 can calculate the variable importance indicating the degree of contribution of the high-priority explanatory variable to the target variable without being affected by the low-priority explanatory variable. That is, the information processing apparatus 100 can calculate the variable importance of the high-priority explanatory variable in a state where the influence of the correlation between the low-priority explanatory variable and the high-priority explanatory variable is excluded. Thereby, the information processing apparatus 100 can accurately estimate, for example, an explanatory variable important for the target variable from among the high-priority explanatory variables based on the variable importance of the high-priority explanatory variable in a state where the influence of the low-priority explanatory variable is excluded. Therefore, the information processing apparatus 100 can accurately estimate an explanatory variable important for the target variable.

[0083] [6. Hardware Configuration] In addition, the information processing apparatus 100 according to the above-described embodiments is realized by, for example, a computer 1000 having a configuration as shown in FIG. 8. FIG. 8 is a hardware configuration diagram showing an example of a computer that realizes the functions of the information processing apparatus. The computer 1000 includes a CPU 1100, a RAM 1200, a ROM 1300, an HDD 1400, a communication interface (I / F) 1500, an input / output interface (I / F) 1600, and a media interface (I / F) 1700.

[0084] The CPU 1100 operates based on programs stored in the ROM 1300 or the HDD 1400, and controls each part. The ROM 1300 stores a boot program executed by the CPU 1100 when the computer 1000 is started up, programs dependent on the hardware of the computer 1000, and the like.

[0085] The HDD 1400 stores programs executed by the CPU 1100, data used by such programs, and the like. The communication interface 1500 receives data from other devices via a predetermined communication network and sends it to the CPU 1100, and transmits data generated by the CPU 1100 to other devices via a predetermined communication network.

[0086] The CPU 1100 controls output devices such as a display and a printer, and input devices 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. In addition, the CPU 1100 outputs generated data to the output device via the input / output interface 1600.

[0087] 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), 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.

[0088] For example, when the computer 1000 functions as the information processing apparatus 100 according to the embodiment, the CPU 1100 of the computer 1000 realizes the functions of the control unit 130 by executing the program loaded onto the RAM 1200. The CPU 1100 of the computer 1000 reads and executes these programs from the recording medium 1800, but as another example, these programs may be acquired from another device via a predetermined communication network.

[0089] As described above, some of the embodiments of the present application have been described in detail with reference to the drawings, but these are examples, and the present invention can be implemented in other forms with various modifications and improvements based on the knowledge of those skilled in the art, including the aspects described in the column of the disclosure of the invention.

[0090] 〔7. Others〕 Also, among the respective processes described in the above embodiments and modification examples, all or part of the processes described as being automatically performed can be manually performed, or all or part of the processes described as being manually performed can be automatically performed by a known method. In addition, regarding the processing procedures, specific names, information including various data and parameters shown in the above documents and drawings, they can be arbitrarily changed unless otherwise specified. For example, the various information shown in each figure is not limited to the illustrated information.

[0091] In addition, each component of each illustrated device is functionally conceptual and does not necessarily have to be physically configured as shown in the figures. That is, the specific form of distribution and integration of each device is not limited to that shown in the figures, and all or part of it can be functionally or physically distributed and integrated in any unit according to various loads, usage situations, etc.

[0092] Also, the above-described embodiments and modifications can be appropriately combined within a range that does not conflict with the processing content.

Description of Reference Numerals

[0093] 100 Information processing device 110 Communication unit 120 Storage unit 130 Control unit 131 Reception unit 132 Generation unit 133 Acquisition unit 134 Provision unit

Claims

1. Based on a plurality of sets of data that are pairs of explanatory data corresponding to each of a plurality of types of explanatory variables and target data corresponding to a target variable, among the plurality of types of explanatory variables, a first decision tree model for predicting the target data from first explanatory data corresponding to each of the first type of explanatory variables is generated, and based on the set of data corresponding to each leaf node of the generated first decision tree model, among the plurality of types of explanatory variables, a second decision tree model for predicting the target data from second explanatory data corresponding to a second type of explanatory variables different from the first type of explanatory variables is generated, a generation unit; An acquisition unit that acquires a second variable importance generated when the generation unit generates the second decision tree model; An information processing apparatus comprising:

2. The acquisition unit: Acquires the second variable importance corresponding to each of the second decision tree models generated by the generation unit, and calculates a total importance, which is the variable importance corresponding to each of the second type of explanatory variables, by summing the acquired second variable importance for each of the second type of explanatory variables. The information processing apparatus according to claim 1.

3. The generation unit: Generates a plurality of different first decision tree models by varying parameters when generating the decision tree model, and generates a plurality of different second decision tree models corresponding to each of the plurality of different first decision tree models generated; The acquisition unit: Calculates an average of the total importance corresponding to each of the plurality of different second decision tree models generated by the generation unit for each of the second type of explanatory variables. The information processing apparatus according to claim 2.

4. The acquisition unit: Obtain the first predicted value by the first decision tree model generated by the generation unit, and verify the prediction accuracy of the first decision tree model based on the first error, which is the error between the verification target data included in the verification set data and the first predicted value. Obtain the second predicted value by the second decision tree model generated by the generation unit, and verify the prediction accuracy of the second decision tree model based on the second error, which is the error between the verification target data and the second predicted value. The information processing apparatus according to claim 1.

5. The acquisition unit Obtain the second variable importance corresponding to the second decision tree model in which the second error is less than the second threshold. The information processing apparatus according to claim 4.

6. The first type of explanatory variable is an explanatory variable corresponding to attribute information regarding the attributes of the subject to be analyzed. The second type of explanatory variable is information regarding the subject and is an explanatory variable corresponding to non-attribute information of a type different from the attributes of the subject. The information processing apparatus according to claim 1.

7. The first type of explanatory variable is an explanatory variable designated as an explanatory variable with a low priority for obtaining variable importance by the user. The second type of explanatory variable is an explanatory variable designated as an explanatory variable with a high priority for obtaining variable importance by the user. The information processing apparatus according to claim 1.

8. An information processing method realized by a program executed by an information processing apparatus, Based on a plurality of set data, which are sets of explanatory data corresponding to each of a plurality of types of explanatory variables and target data corresponding to a target variable, a first decision tree model for predicting the target data from first explanatory data corresponding to each of the first type of explanatory variables among the plurality of types of explanatory variables is generated. Based on the set data corresponding to each of the leaf nodes of the generated first decision tree model, a second decision tree model for predicting the target data from second explanatory data corresponding to a second type of explanatory variables different from the first type of explanatory variables among the plurality of types of explanatory variables is generated. A generation step; An acquisition step of acquiring a second variable importance generated when the generation step generates the second decision tree model; An information processing method including the above.

Images