Information processing device, information processing method and information processing program

Abstract

An information processing device (1) includes a determination unit (33) and a presentation unit (34). In the case of calculating an intervention effect in a target variable produced by intervention on any one of a plurality of variables based on causal information (41) indicating a causal relationship between a plurality of variables, the determination unit (33) determines whether data of the variable required for the calculation is insufficient. The presentation unit (34) presents information based on the determination result of the determination unit (33) to a user.

Description

Technical Field

[0001] The present disclosure relates to an information processing device, an information processing method, and an information processing program. Background Art

[0002] Conventionally, there are deep learning, multivariate regression, decision tree, random forest, etc. as a technique for predicting the value of a variable as a target from other variables in multivariate data analysis. In addition to such a technique for performing prediction by using the correlation between variables, data analysis based on a so-called causal graph generated by a technique for inferring a causal relationship between variables has been proposed (see, for example, Patent Document 1).

[0003] Citation List

[0004] Patent Literature

[0005] Patent Document 1: JP 2014-228991 A Summary of the invention

[0006] Technical issues

[0007] Incidentally, there is a technique for calculating the causal effect on a variable as a target by intentionally changing the value of the variable (intervention variable) when a person using data (such as a company) performs data analysis (intervention effect calculation). In this intervention effect calculation, in the case where there is an infinite (or sufficient) amount of data for each variable, a highly accurate calculation result can be output. However, in reality, since the data for each variable is limited and the number of data items for each variable varies, there is a possibility that the accuracy of the calculation result is reduced. Therefore, in the case of presenting a low-precision calculation result, it cannot be said that appropriate information is presented to the user, and there is room for improvement.

[0008] Therefore, the present disclosure proposes an information processing device, an information processing method, and an information processing program, which are capable of presenting appropriate information to a user according to the state of data for performing intervention effect calculation.

[0009] Problem Solution

[0010] An information processing device includes a determination unit and a presentation unit. In the case of calculating an intervention effect in a target variable produced by intervention of any one of a plurality of variables based on causal information indicating a causal relationship between a plurality of variables, the determination unit determines whether data of the variable required for the calculation is insufficient. The presentation unit presents information based on the determination result of the determination unit to a user. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 is a diagram showing an overview of an information processing method according to an embodiment of the present disclosure.

[0012] Figure 2 is a block diagram showing the configuration of an information processing apparatus according to the present embodiment.

[0013] Figure 3 is a diagram showing an example of customer information.

[0014] Figure 4 is a diagram showing information presented by a presentation unit.

[0015] Figure 5 is a diagram showing information presented by a presentation unit.

[0016] Figure 6 is a diagram showing information presented by a presentation unit.

[0017] Figure 7 is a diagram showing information presented by a presentation unit.

[0018] Figure 8 is a diagram showing information presented by a presentation unit.

[0019] Fig. 9 is a diagram showing information presented by a presentation unit.

[0020] Fig.10 is a diagram showing information presented by a presentation unit.

[0021] Fig.11 is a diagram showing information presented by a presentation unit.

[0022] Fig.12 is a diagram showing information presented by a presentation unit.

[0023] Fig.13 is a diagram showing information presented by a presentation unit.

[0024] Fig.14 is a diagram showing information presented by a presentation unit.

[0025] Fig.15 is a diagram showing information presented by a presentation unit.

[0026] Fig.16 is a diagram showing information presented by a presentation unit.

[0027] Fig.17 is a flowchart showing an information processing procedure performed by the information processing apparatus according to the present embodiment.

[0028] Fig.18 is a block diagram showing an example of the hardware configuration of the information processing apparatus according to the present embodiment. DETAILED DESCRIPTION

[0029] Hereinafter, embodiments of the present disclosure will be described in detail based on the drawings. Note that in each of the following embodiments, duplicate descriptions are omitted by assigning the same reference numerals to the same parts.

[0030] In addition, in the present specification and the drawings, a plurality of components having substantially the same functional configuration may be distinguished by assigning different numerals after the same reference numerals. However, in the absence of a particular need to distinguish a plurality of components having substantially the same functional configuration from each other, only the same reference numerals are assigned.

[0031] In addition, for example, an example of data analysis in a company (such as a service company) that develops products or services to provide customers will be described in the following embodiments. Note that the information processing device, information processing method, and information processing program according to the present embodiment are not limited to the case of being applied to data analysis of a company (i.e., data analysis in the business field), and can be applied to data analysis in various fields (e.g., the medical field and the educational field). In other words, the information processing device, information processing method, and information processing program according to the present embodiment can be applied to data analysis that processes multiple variables.

[0032] <<1. Overview of Information Processing Method According to Present Embodiment>>

[0033] First, refer to Figure 1 An outline of the information processing method according to the present embodiment is described. Figure 1 is a diagram showing an overview of an information processing method according to an embodiment of the present disclosure. Figure 1 The causal information indicating the causal relationship between multiple variables, the so-called causal diagram, is shown in FIG. Figure 1 In the example shown, the direction of the causal relationship between variables is indicated by the arrows in the causal diagram (cause → effect). That is, Figure 1 The causal graph shown is a directed graph. In addition, Figure 1 Causal information as shown in is in other words information with a graphical model of probability distribution where variables of probabilistic / statistical causes and effects are connected by arrows.

[0034] Note that any variable such as a categorical variable or a continuous variable can be used. However, it is preferred that the variable is specifically a categorical variable. In the case of a continuous variable, a categorical variable can be obtained by n-equal divisions of a distribution function (n is a natural number of 2 or greater), etc.

[0035] also, Figure 1 The causal diagram shown is an example of causal information. The causal information only needs to be information that can grasp the causal relationship between multiple variables, and can be information that lists the causal relationship between variables.

[0036] For example, such causal information is generated using attribute data (such as age, gender, or address) of customers whose data is owned by the company, data of a questionnaire conducted on each customer, and the like.

[0037] For example, in Figure 1 , causal information generated by using data from companies that provide services A and B to customers who are (current and past) members is shown. Specifically, Figure 1 The variable Z1 shown in is the address data of the customer who is a member. In addition, the variable Z2 is the questionnaire data indicating the satisfaction with the service B. In addition, the variable X is the questionnaire data indicating the satisfaction with the service A. In addition, the variable Y is the data indicating whether the customer has withdrawn from the service.

[0038] Note that service A and service B are not limited to being provided by the same company, but may be provided by different companies. That is, causal information (causal graph) is not limited to being generated using data from one company, but may be generated using data from multiple companies.

[0039] Here, for example, it is assumed that the company wants to see how much variable Y changes when variable X in causal information is intentionally changed in order to reduce the dropout rate of customers. That is, it is assumed that calculation of the intervention effect generated in variable Y as a target when variable X is intervened (hereinafter, intervention effect calculation or intervention calculation) is performed. Note that, hereinafter, variable X to be intervened may be referred to as intervention variable X, and variable Y to be a target may be referred to as target variable Y.

[0040] Traditionally, in such intervention effect calculations, when the amount of data for each variable is infinite (or sufficient), a highly accurate calculation result can be output. However, in reality, the data for each variable is limited, and there are also cases where the number of data items for the variable changes. In this case, the accuracy of the calculation result may be reduced. That is, traditionally, it cannot be considered that appropriate information is presented to the user because even a calculation result with low accuracy is presented to the user.

[0041] Therefore, in the information processing method according to the present embodiment, information corresponding to the data state for performing intervention effect calculation is presented to the user. Figure 1 Give a description.

[0042] Note that assuming Figure 1 The causal information shown is pre-generated based on the company's customer data. Figure 1 In the example, it is assumed that a variable X (intervention variable X) and a target variable Y of the intervention effect that is expected to be seen are received from the user, where the user (such as the person in charge of a company, etc.) wants to intervene in the variable X. That is, in Figure 1In the example, suppose an intervention is conducted on the data on satisfaction with service A, and the change in whether members withdraw or do not withdraw is considered the intervention effect.

[0043] like Figure 1 As shown, in the information processing method according to the present embodiment, when calculating the intervention effect produced in the target variable Y by intervening in any variable (intervention variable X) among multiple variables X, Y, Z1 and Z2 based on causal information, it is determined whether the data of variables X, Y, Z1 and Z2 required for the calculation of the intervention effect are insufficient (step S1).

[0044] The data of the variables mentioned herein include data obtained by combination of a plurality of variables in addition to the data of each variable, and the details thereof will be described later.

[0045] Then, in the information processing method according to the present embodiment, information based on the determination result in step S1 is presented to the user (step S2 ).

[0046] Specifically, in the information processing method according to this embodiment, when it is determined that there is insufficient data (sufficient data) for the variables X, Y, Z1 and Z2 required for calculating the intervention effect, the intervention effect is calculated based on the intervention content received from the user, and the calculation results are presented to the user.

[0047] On the other hand, in the information processing method according to the present embodiment, in the case where it is determined that the data of the variables X, Y, Z1, and Z2 required for the intervention effect calculation are insufficient, information indicating the impossibility of the intervention effect calculation, information indicating the variables for which the data are insufficient, information indicating proposed measures for satisfying the data of the variables required for the intervention effect calculation, etc. are presented to the user. Note that the details of the information to be presented to the user will be described later.

[0048] That is, in the information processing method according to the present embodiment, in the case of insufficient data, a notification is given indicating that high-precision intervention effect calculation cannot be performed using the current data, or proposed measures that can achieve high-precision intervention effect calculation are presented to the user instead of performing the intervention effect calculation.

[0049] As a result, it is possible to prevent the presentation of a calculation result with low precision without the user's intention, and it is also possible to inform the user that the data is insufficient under the current number of data items. Therefore, in the information processing method according to this embodiment, appropriate information can be presented to the user according to the data state for performing the intervention effect calculation.

[0050] <<2. Configuration of Information Processing Device According to the Present Embodiment>>

[0051] Next, we will refer to Figure 2 The configuration of the information processing apparatus 1 according to the present embodiment is described. Figure 2 1 is a block diagram showing the configuration of the information processing device 1 according to the present embodiment. Figure 2 As shown, the information processing device 1 is communicably connected to the user terminal 11 via a predetermined network (not shown). Note that in the present embodiment, a case where the information processing device 1 and the user terminal 11 are configured separately is described. However, in another embodiment, a terminal device configured in a manner that integrates the functions of the information processing device 1 and the user terminal 11 may be adopted.

[0052] The user terminal 11 is a terminal device used by a user such as a company manager or an individual. For example, the user terminal 11 is implemented by a smartphone, a tablet terminal, a notebook personal computer (PC), a desktop PC, a cellular phone, a personal digital assistant (PDA), etc.

[0053] In addition, the information processing device 1 includes a communication unit 2, a control unit 3, and a storage unit 4. The communication unit 2 is implemented by, for example, a network interface card (NIC), etc. Then, the communication unit 2 transmits and receives information to and from the user terminal 11 via a predetermined network.

[0054] The control unit 3 includes a receiving unit 31, an extracting unit 32, a determining unit 33, a presenting unit 34, a proposed measure executing unit 35, and an intervention calculating unit 36. The storage unit 4 stores causal information 41 and customer information 42.

[0055] Here, the information processing device 1 includes, for example, a computer having a central processing unit (CPU), a read only memory (ROM), a random access memory (RAM), a data flash memory, an input / output port, etc., and various circuits.

[0056] The CPU of the computer functions as the receiving unit 31 , extracting unit 32 , determining unit 33 , presenting unit 34 , proposed measure executing unit 35 , and intervention calculating unit 36 ​​of the control unit 3 , for example, by reading and executing programs stored in the ROM.

[0057] In addition, at least one or all of the receiving unit 31, extraction unit 32, determination unit 33, presentation unit 34, proposed measure execution unit 35 and intervention calculation unit 36 ​​of the control unit 3 can be configured by hardware, such as an application-specific integrated circuit (ASIC) or a field programmable gate array (FPGA).

[0058] In addition, the storage unit 4 corresponds to, for example, a RAM and a data flash memory. The RAM and the data flash memory can store causal information 41, customer information 42, information of various programs, etc. Note that the information processing device 1 can obtain the above-mentioned programs and various information through another computer or a portable recording medium connected through a wired or wireless network.

[0059] like Figure 1 As shown, the causal information 41 is information indicating a probability or statistical causal relationship between a plurality of variables. For example, the causal information 41 may be created based on a statistical estimation model of a causal Bayesian network or a causal structural equation (see, for example, Patent Document 1), or may be stored as information in which an expert or user associates causal relationships between variables.

[0060] The customer information 42 is data of customers whose data the company to which the user belongs has. Figure 3 is a diagram showing an example of the customer information 42. Note that the customer information 42 may be generated for each company, or data of customers whose data a plurality of companies have may be integrated.

[0061] like Figure 3 As shown, the customer information 42 includes items such as "customer ID", "age", "sex", "address", and "questionnaire data".

[0062] "Customer ID" is identification information for identifying a customer. "Age" is information indicating the age of the customer. Note that "age" can be Figure 3 The exact age shown, or it can be abstracted as "twenties". "Gender" is information indicating the gender of the customer. "Address" is information indicating the address of the customer. "Address" can be a detailed address, or it can be abstracted as "Tokyo" or "Kanto area".

[0063] "Questionnaire data" is information indicating answers to a questionnaire administered by a company to a customer. Note that Figure 3 "Not answered" in the "questionnaire data" shown indicates that the questionnaire was not executed, or that although the questionnaire was executed, no answer was given.

[0064] Next, the respective functional blocks (the receiving unit 31 , the extracting unit 32 , the determining unit 33 , the presenting unit 34 , the proposed measure executing unit 35 , and the intervention calculating unit 36 ​​) of the control unit 3 will be described.

[0065] The receiving unit 31 receives various information from the user via the user terminal 11. For example, the receiving unit 31 receives information for performing intervention effect calculation from the user. Specifically, the receiving unit 31 receives the selection of an intervention variable and a target variable from the variables in the causal information 41 from the user, the intervention variable being the variable to be intervened, and the target variable being the variable on which the intervention effect is expected to be seen. In addition, the receiving unit 31 receives the user's selection operation of the options presented from the information processing device 1. Specifically, in the case where there is missing data in the variables extracted by the extraction unit 32 (described later), the receiving unit 31 receives a selection operation to specify processing for the missing variables. In addition, in the case where the intervention effect calculation cannot be performed, the receiving unit 31 receives a selection operation to specify a proposed measure to be performed by the proposed measure execution unit 35 (described later) from a plurality of proposed measures presented to the user. In addition, in the case where the intervention effect calculation can be performed, the receiving unit 31 receives the intervention content (changed content) regarding the intervention variable.

[0066] In the case where the intervention variable and the target variable are received by the receiving unit 31, the extraction unit 32 extracts the variables required for the intervention effect calculation from the variables included in the causal information 41. Specifically, the extraction unit 32 extracts the variables (such as the confounding variables) that have a direct or indirect causal relationship with the intervention variable based on the causal information 41. In other words, the extraction unit 32 extracts Figure 1 The causal arrows shown in are connected to the variables that intervene in variable X (variables Z1 and Z2) and the causal arrows are connected to the variables (variables Z1 and Z2) but not to the variables that intervene in variable X.

[0067] In the case of performing intervention effect calculation based on causal information 41, the determination unit 33 determines whether the data of the variables required for the intervention effect calculation are insufficient. First, before determining that the data are insufficient, the determination unit 33 determines whether the variables extracted by the extraction unit 32, the intervention variables, and the target variables contain variables with missing data. Variables with missing data indicate a situation where data for a part of the possible values ​​of a certain variable do not exist. For example, for frequency distribution data of a continuous variable, it means that data for a part of the values ​​(levels) of the continuous values ​​(levels) do not exist (are missing).

[0068] When there is a variable with missing data (hereinafter referred to as a missing variable), the determination unit 33 presents the method of processing the missing variable to the user and receives the user's selection. For example, the determination unit 33 presents the following three as methods of processing the missing variable.

[0069] (1) Excluding data with missing values

[0070] (2) Filling in missing values

[0071] (3) Treat missing values ​​as categorical values

[0072] (1) Excluding data with missing values

[0073] In the case where “Exclude data with missing values” is selected, the determination unit 33 performs aggregation processing at the next stage based on the assumption that there is no data for the missing data in the missing variable.

[0074] (2) Filling in missing values

[0075] In the case where "Supplement missing values" is selected, the determination unit 33 supplements the missing data in the missing variable. For example, in the case where the missing variable is a variable having continuous values ​​(continuous variable), the determination unit 33 supplements the data of the missing value by using, for example, the average value or median value of the data of other values ​​included in the missing variable. In addition, in the case where the missing variable is a categorical variable, the determination unit 33 supplements the data of the missing value by using, for example, a representative value in the missing variable.

[0076] (3) Treat missing values ​​as categorical values

[0077] In the case where "treat missing value as a categorical value" is selected, the determination unit 33 processes the missing data as it is. Specifically, the determination unit 33 adds information indicating the presence of missing to the missing value in the missing variable, and performs aggregation processing in the next stage.

[0078] In the case where the data of the variables are not missing, and in the case where the processing corresponding to the method of processing missing variables selected by the user is completed, the determination unit 33 aggregates the data of these variables. Specifically, the determination unit 33 performs an aggregation process of aggregating data for each combination of the intervention variable, the target variable, and the variable extracted by the extraction unit 32.

[0079] Then, the determination unit 33 determines whether the number of data items is equal to or greater than a predetermined threshold value for each combination of variables for which aggregation is performed by the aggregation process. In the case where the number of data items for all combinations of variables is equal to or greater than the threshold value, the determination unit 33 determines that the data of the variables required for the intervention effect calculation is sufficient (not insufficient). That is, the determination unit 33 determines that the intervention effect calculation can be performed.

[0080] On the other hand, if there is a combination of variables with a number of data items less than the threshold, the determination unit 33 determines that the data of the variables required for intervention effect calculation is insufficient. In other words, the determination unit 33 determines that intervention effect calculation cannot be performed.

[0081] The determination unit 33 notifies the determination result to the presentation unit 34 and the intervention calculation unit 36. Note that the determination result presented to the presentation unit 34 includes information indicating whether data of variables required for intervention effect calculation is insufficient, information on combinations of variables having insufficient data, and the like.

[0082] In addition, when the proposed measure is executed by the proposed measure execution unit 35, the determination unit 33 determines whether the data of the variables required for intervention effect calculation is insufficient. That is, the determination unit 33 determines whether the insufficient data is resolved by executing the proposed measure by the proposed measure execution unit 35.

[0083] The presentation unit 34 presents information to the user based on the determination result of the determination unit 33. For example, when the determination unit 33 determines that the data of the variables required for the intervention effect calculation is sufficient, the presentation unit 34 presents information indicating that the intervention effect calculation can be performed to the user.

[0084] On the other hand, in the case where the determination unit 33 determines that the data of the variables required for the intervention effect calculation is insufficient, the presentation unit 34 presents to the user information indicating that the intervention effect calculation cannot be performed. In addition, the presentation unit 34 presents the information of the insufficient data together with the proposed measures for satisfying the data and the information indicating that the intervention effect calculation cannot be performed. Note that reference will be made later on. Figures 4 to 16 Details of the information presented by the presentation unit 34 are described.

[0085] The proposed measure execution unit 35 executes the proposed measure presented to the user by the presentation unit 34. For example, in the case where a plurality of proposed measures are presented by the presentation unit 34, the reception unit 31 receives a selection of one proposed measure from the user, and further receives an execution instruction instructing to execute the selected proposed measure. Then, in the case where the reception unit 31 receives an execution instruction for the proposed measure, the proposed measure execution unit 35 executes the proposed measure. Note that reference will be made later to Figures 4 to 16 A specific example of the proposed measure executed by the proposed measure execution unit 35 is described.

[0086] When the determination unit 33 determines that the data of the variables required for intervention effect calculation are sufficient, the intervention calculation unit 36 ​​performs intervention effect calculation. Specifically, the intervention calculation unit 36 ​​calculates the intervention effect in the target variable based on the intervention content on the intervention variable received by the reception unit 31 from the user.

[0087] Then, the intervention calculation unit 36 ​​notifies the presentation unit 34 of the calculation result of the intervention effect calculation, and the presentation unit 34 presents intervention information based on the calculation result to the user. Note that the details of the intervention information presented to the user will be described later.

[0088] <<3. Specific examples of information presented to users>>

[0089] Next, we will refer to Figures 4 to 16 Details of the information presented to the user by the presentation unit 34 are described.

[0090] Figures 4 to 16 is a diagram showing information presented by the presentation unit 34 .

[0091] For example, Figure 4 As shown, the presentation unit 34 presents information indicating that the intervention effect calculation is not feasible, such as “Error!” and “Cannot perform intervention calculation”, and presents information indicating that the variable has insufficient data, such as “The combined data of address = “Kanto” and satisfaction with service A = “4: Slightly satisfied” is insufficient”. Therefore, the user can accurately grasp the reason why the intervention effect calculation cannot be performed.

[0092] Note that, as information to be presented to the user, the presentation unit 34 may only display Figure 4 The information (text information) shown in the lower part can also be displayed Figure 4 The information shown in the upper part (graphic model of causal information 41).

[0093] Then, for example, Figure 5 As shown, the presentation unit 34 may present information indicating that intervention effect calculation is not feasible, and present proposed measures to satisfy data of variables required for intervention effect calculation.

[0094] exist Figure 5 The following five proposed measures are presented and check boxes are displayed. Note that reference will be made later. Figures 7 to 16 Describe the specific content of proposed measures (1) to (5).

[0095] (1) Classification value of binding variable ○

[0096] (2) Remove Zi with small influence

[0097] (3) Change the direction of the arrow from X to Zi

[0098] (4) Continuing the calculation with reduced precision

[0099] (5) Collect data again

[0100] Note that "variable ○" is an arbitrary variable related to the calculation of the intervention effect and is Figure 4 In the example of , it is variable X, variable Z1 or variable Z2. In addition, it is assumed that "Z1" is variable Z1 or variable Z2 hereinafter.

[0101] In addition, if Figure 5As shown, in the case where there is a proposed measure that does not resolve the data shortage even when executed as a proposed measure, the presentation unit 34 slightly displays the proposed measure ( Figure 5 (2) Remove Zi with small influence). That is, the presentation unit 34 changes the display mode of the proposed measures that cannot solve the insufficient data, compared with the display mode of other proposed measures. Therefore, it is possible to prevent the user from mistakenly selecting meaningless proposed measures. Note that the proposed measures that cannot solve the insufficient data can be hidden.

[0102] In addition, if Figure 5 As shown, in the case where a plurality of proposed measures are presented, the presenting unit 34 presents a plurality of proposed measures based on the user's proficiency level ( Figure 5 Note that the user's proficiency level indicates the level of experience in data analysis.

[0103] For example, Figure 6 As shown in FIG. 1 , in the case where the user's proficiency is high (experienced), the presentation unit 34 makes the recommended information of "changing the direction of the arrow from X→Zi" high. That is, for the proposed measure "changing the direction of the arrow from X→Zi", the presentation unit 34 dynamically changes the degree of the recommended information according to the proficiency, because when the proposed measure is performed by a user with a low proficiency, the reliability of the intervention effect calculation decreases. In addition, as Figure 6 As shown, the presentation unit 34 displays auxiliary information such as “Since you are an expert, the recommendation degree is set higher than usual.” Therefore, in the case of executing the proposed measure “Change the direction of the arrow from X→Zi,” it can be notified that a high proficiency is required.

[0104] Then, the receiving unit 31 receives a selection (checking of a check box) of at least one proposed measure from the user. Subsequently, the presenting unit 34 presents the proposed measure received by the receiving unit 31 as follows: Figures 7 to 14 The specific information shown. In the following, the above-mentioned proposed measures (1) to (5) will be described in detail.

[0105] <(1) Classification value of binding variable ○>

[0106] exist Figure 7 A specific example of the proposed measure "binding the classification value of variable ○" is shown in FIG. Figure 7 In the example of , it is assumed that X among X, Z1, and Z2 as candidates for variable ○ is selected by determination of the system or the user. In the case where the user selects the proposed measure “bind classification value of variable ○”, the presentation unit 34 binds data of a value having insufficient data with data of another value among a plurality of values ​​(classification values) included in the variable ○.

[0107] exist Figure 7, the distribution of a plurality of pieces of data (the number of customers) at each value of “satisfaction with service A” as the intervention variable X is shown. Note that the values ​​(1 to 5) of “satisfaction with service A” respectively indicate “1: very dissatisfied”, “2: slightly dissatisfied”, “3: neither”, “4: slightly satisfied”, and “5: very satisfied”.

[0108] Here, for example, it is assumed that the number of data items of "1: Very dissatisfied" is less than a predetermined threshold value. In this case, for example, the presentation unit 34 presents to bind multiple data items "1: Very dissatisfied" and "2: Slightly dissatisfied". That is, the presentation unit 34 presents to regard the following two values ​​as one value: "1: Very dissatisfied" for a value with insufficient number of data items, and "2: Slightly dissatisfied" for a value with sufficient number of data items, and add the number of data items. That is, it is proposed to regard two values ​​with relatively similar contents as one value, and add the number of data items.

[0109] Then, in the case where the receiving unit 31 receives an execution instruction for the proposed measure from the user, the proposed measure execution unit 35 executes the proposed measure presented by the presentation unit 34. That is, the proposed measure execution unit 35 binds "1: Very dissatisfied" and "2: Slightly dissatisfied", and adds the number of data pieces of the values. Note that the binding value in this case is preferably expressed as a value that can identify that "1: Very dissatisfied" and "2: Slightly dissatisfied" are bound.

[0110] As a result, since the number of data items of the multiple values ​​considered as one value is the sum of the number of data items of the two values, the number of data items is not insufficient. In other words, by adding the number of data items of the two values, the possible values ​​of the variable X are reduced from 5 values ​​to 4 values, solving the problem of insufficient data. Therefore, the intervention effect calculation can be performed by the intervention calculation unit 36 ​​(described later).

[0111] Note that the case of binding variable values ​​is not limited to the case where a value with insufficient data and a value with sufficient data are bound. As long as the insufficient data can be solved, a value with insufficient data can be bound. Figure 7 , a case where two values ​​are bound is shown, but three or more values ​​may be bound.

[0112] In addition, although Figure 7 , the binding process is shown in the case where the number of data items of one variable X is insufficient, but the binding process can be performed even when the data of a combination of multiple variables is insufficient. Figure 8 Give a description.

[0113] exist Figure 8 In , it is assumed that the data for the combination of the variable "address" and the variable "satisfaction with service A" is insufficient. Figure 8As shown, in the case where the data of a combination of multiple variables is insufficient, the presentation unit 34 visualizes the information indicating the variables with insufficient data and presents it to the user. Specifically, the presentation unit 34 displays a list of the number of data items of the variable combination including the variables with insufficient data in a table format. In addition, according to the number of data items, the presentation unit 34 changes the color information (shading or RGB) of the background of each item displayed in the table format. As a result, the user can easily grasp the data of the insufficient combination of values ​​in the multiple variables. Note that the presentation unit 34 is not limited to the case of changing the color information of the background of each item, and only needs to be able to change the display mode (such as character size) according to the number of data items of each item.

[0114] Then, the presentation unit 34 presents the binding of the variable value and the information in the above table. Figure 8 , the presentation unit 34 presents two binding methods to solve the insufficient data of the combination of "Hokkaido" and "very dissatisfied". Specifically, the presentation unit 34 presents a method of binding "Hokkaido" and "Tohoku" and a method of binding "very dissatisfied" and "slightly dissatisfied". That is, the presentation unit 34 solves the insufficient data by binding the value of the variable "address" and reducing the value from 5 to 4, or binding the value of the variable "satisfaction with service A" and reducing the value from 5 to 4.

[0115] Then, in the case where the receiving unit 31 receives an instruction to execute any one of the proposed measures from the user, the proposed measure executing unit 35 executes the selected proposed measure.

[0116] In addition, if Fig. 9 As shown, the presentation unit 34 performs binding processing internally and does not present it to the user in such a way that areas with a small number of data entries (individual items of the combination of variables) are eliminated, and clustering processing is performed on the values ​​(classification values) of variables X and variables Z of each variable or the combination patterns of the variables in the internal binding processing, and the conditions for performing intervention effect calculation are met.

[0117] Specifically, Figure 8 As shown, the presentation unit 34 internally performs a binding process of eliminating areas where the number of data items is insufficient by clustering items having similar distributions of values ​​of the variables. More specifically, in Figure 8 In the example shown, the presentation unit 34 performs binding with a combination of variable X and variable Z.

[0118] <(2) Removing Zi with small influence>

[0119] Next, in Fig.10A specific example of the proposed measure "remove Zi with small influence" is shown in . In the case where the proposed measure "remove Zi with small influence" is selected by the user, the presentation unit 34 presents to the user the variables having small influence on the intervention effect among the variables required for the intervention effect calculation, which are excluded from the intervention effect calculation.

[0120] Here, in Fig.10 In the figure, the arrows indicating the causal relationship between the variables are displayed thicker as the degree of influence on the intervention effect becomes larger. Note that the degree of influence on the intervention effect can be calculated, for example, based on the change in the mutual information (the amount of mutual dependence between the two variables) between the intervention variable X and the target variable Y. Specifically, the above-mentioned degree of influence can be calculated as the difference between the mutual information between the intervention variable X and the target variable Y when each variable Z1 (or variable Z2) is adjusted and the mutual information when each variable is not adjusted. That is, the smaller the difference, the smaller the degree of influence.

[0121] exist Fig.10 In the example shown, variable Z1 "address" has little influence on the intervention effect calculation when setting the intervention variable X and the target variable Y. In other words, even when variable Z1 "address" is removed from the intervention effect calculation, the reliability of the calculation result does not decrease much.

[0122] Therefore, the presentation unit 34 presents to the user the removal of the variable Z1 from the intervention effect calculation. As a result, for example, in the case where the reason why the intervention effect calculation cannot be performed is insufficient data of the variable Z1, by removing the variable Z1 by the proposed measure execution unit 35, the intervention effect calculation can be performed while minimizing the decrease in the reliability of the calculation result.

[0123] Note that the presentation unit 34 may present only the variables with the smallest degree of influence as the variables to be removed from the intervention effect calculation, or may display a list of all variables in order of degree of influence and allow the user to select the variables to be removed from the calculation. Alternatively, the presentation unit 34 may present variables with a degree of influence less than a predetermined threshold, that is, variables with a difference between the above mutual information amounts less than a predetermined threshold. In addition, for example, in the case where the mutual information amount between the variable X and the variable Z1 or the mutual information amount between the variable Y and the variable Z1 is less than a predetermined threshold, the presentation unit 34 may present the variables to be removed from the calculation.

[0124] Note that the above-mentioned degree of influence is not limited to the case where it is calculated based on the mutual information amount of two variables, and can be calculated based on, for example, another method such as a function that calculates the correlation between two variables, a function that calculates the similarity between two variables, or a propensity score. That is, the calculation can be performed by any method that measures the distance between two variables.

[0125] Note that for the proposed measure "remove Zi with small influence", when the above-mentioned influence degree is less than a predetermined threshold and there is no influence even when removal from the calculation is performed, the presentation unit 34 can internally perform the processing of removing from the calculation without presenting the proposed measure to the user.

[0126] <(3) Change the direction of the arrow from X to Zi>

[0127] Next, in Fig.11 A specific example of the proposed measure “change the direction of the arrow from X→Zi” is shown in . When the user selects the proposed measure “change the direction of the arrow from X→Zi”, the presentation unit 34 presents the user with the direction of changing the cause and effect between the variables required for intervention effect calculation.

[0128] exist Fig.11 Variable Z2 → variable X before the change is changed to variable X → variable Z1 after the change. Note that the user can arbitrarily select the variables whose directions of cause and effect are to be changed, or the control unit 3 can automatically select two variables with high correlation, high similarity, etc.

[0129] In addition, if Fig.11 As shown, the presentation unit 34 presents to the user proposed measures to change the direction of the cause and the result, and presents supplementary information related to the proposed measures. Fig.10 In the example shown, the presentation unit 34 displays “tips” and suggested measures as supplementary information.

[0130] For example, Fig.11 As shown, in "Hint" as supplementary information, information related to a specific example of a situation in which the direction of cause and effect between two variables can be changed is presented.

[0131] Then, in the case where the receiving unit 31 receives an execution instruction for the proposed measure from the user, the proposed measure execution unit 35 executes the proposed measure. As a result, since the direction of cause and effect between the two variables is reversed, the variables required for the intervention effect calculation are changed, such as Fig.12 shown.

[0132] exist Fig.12 , which shows the case where the direction of cause and effect between variable X and variable Z2 has changed. For example, in the case of setting intervention variable X and target variable Y, variable Z1 and variable Z2 are variables required for intervention effect calculation in the causal relationship before the change, which is in Fig.12 is shown in the upper part of .

[0133] Here, if the data of variable Z2 is insufficient, it cannot be used Fig.12The causal relationship shown in the upper part of performs the intervention effect calculation. Therefore, if Fig.12 As shown in the lower part of , by changing the direction of cause and effect between variables X and Z2, variable Z2 is no longer a required variable for intervention effect calculation. Fig.12 By using the cause-effect relationship shown in the lower part of , it is possible to perform intervention effect calculation while setting the intervention variable X and the target variable Y. In other words, since the variables required for intervention effect calculation are changed (reduced) by the change in the direction of cause and effect, the variables with insufficient data become unnecessary for calculation, and as a result, insufficient data can be resolved.

[0134] Moreover, in Fig.12 In the example shown, it has been proposed that the direction of cause and effect is changed due to the high correlation or similarity between variable X and variable Z2. However, in the case where the correlation or similarity between variable X and variable Z2 is high, it may be presented that the variable to be intervened is changed from variable X to variable Z2. This point will be referred to Fig.13 Give a description.

[0135] For example, Fig.13 As shown, in the case where the correlation (or the above-mentioned degree of influence) between the variable X and the variable Z2 is equal to or greater than the predetermined threshold, the presentation unit 34 presents changing the variable to be intervened from the variable X to the variable Z2. In other words, in the case where there is a variable having a high correlation with the intervention variable before the change, the presentation unit 34 presents changing the variable to be intervened to the variable. That is, as a proposed measure, the presentation unit 34 presents to the user changing the variable to be intervened without changing the target variable Y.

[0136] Therefore, for example, when the data of variable Z1 is insufficient, variable Z2 is set as the intervention variable, and variable Z1 is no longer a variable required for intervention effect calculation. Insufficient data of variables required for intervention effect calculation can be solved.

[0137] In addition, if Fig.13 As shown, the presentation unit 34 presents proposed measures to the user to change the intervention variable, and presents supplementary information related to the proposed measures. Fig.13 In the example shown, the presentation unit 34 displays a “hint” as supplementary information together with the proposed measure.

[0138] For example, Fig.13 As shown, information related to specific examples of situations in which the intervention variables can be changed is presented in "tips" as supplementary information.

[0139] <(4) Continuing calculation with reduced precision>

[0140] Next, in Fig.14A specific example of the proposed measure “continue calculation when accuracy decreases” is shown in FIG. That is, when the user selects the proposed measure “continue calculation when accuracy decreases”, the presentation unit 34 presents the user with the option to continue the intervention effect calculation and the accuracy of the intervention effect calculation ( Fig.14 The reliability shown in the lower part of ) is reduced.

[0141] Specifically, Fig.14 As shown in the upper part of FIG. 1 , first, the presentation unit 34 displays information for the user to input the intervention content regarding the intervention variable X. Specifically, the presentation unit 34 displays the data of the variable X before the intervention and the data of the variable X after the intervention according to the user operation. That is, the user performs an operation to change the data distribution of the variable X after the intervention. Note that Fig.14 As shown, in the value of the variable X, the display mode of the value intervened by the user is preferably changed relative to the display mode of the value not intervened. Fig.14 In the figure, the data of the user intervention value are indicated by shading.

[0142] Then, in the case where the user operates the “enter” button after the intervention operation, the intervention calculation unit 36 ​​performs intervention effect calculation, and the presentation unit 34 displays intervention information based on the calculation result of the intervention calculation unit 36. Fig.14 As shown in the lower part of , the intervention information presented by the presentation unit 34 includes data of the target variable Y before and after the intervention, and information related to the change in the ratio of the number of data items (intervention effect) (the ratio of "1" increased by ○○%!).

[0143] In addition, if Fig.14 As shown in the lower part of , the presentation unit 34 presents intervention information including the reliability of the intervention effect to the user, such as "However, since the approximate calculation is performed, the reliability of the result is □□%". In addition, preferably, the reliability of the intervention effect is presented by the presentation unit 34 only in the case where the determination unit 33 determines that the data is insufficient to perform the intervention effect calculation originally. That is, in the case where the determination unit 33 determines that the data of the variables required for the intervention effect calculation are sufficient, the presentation unit 34 does not present the information indicating the reliability of the intervention effect.

[0144] Note that in the case of calculating the above reliability (accuracy of intervention effect calculation), the intervention calculation unit 36, for example, uses the data of the variables required for the intervention effect calculation to repeatedly perform random sampling of data → intervention effect calculation, and calculates the reliability by calculating the variance of the calculation results. Alternatively, the intervention calculation unit 36 ​​may calculate a standardized index corresponding to the number of data pieces for each combination of the variables required for the intervention effect calculation, and calculate its minimum value as the reliability.

[0145] Note that the presentation content when continuing the intervention effect calculation when the determination unit 33 determines that the data required for the intervention effect calculation is insufficient is as follows: Fig.14 However, for example, in the case where the determination unit 33 determines that the data required for the intervention effect calculation is sufficient, in addition to presenting the reliability, the same Fig.14 Information that is basically similar to the information in .

[0146] In addition, for example, when the user is allowed to change the distribution of the intervention variable, the user interface (UI) may be restricted in such a way that the user cannot intervene in the value having a small number of data items. Fig.15 Give a description.

[0147] Fig.15 is a diagram showing information presented by the presentation unit 34. Fig.15 In the example, it is assumed that the number of data items with a value of "1" is less than a predetermined number. In this case, when displaying the distribution of data of the variable X after intervention, the presentation unit 34 superimposes the display of "non-keyable" on the data with a value of "1" so that the data with a value of "1" cannot be changed.

[0148] That is, when the intervention calculation unit 36 ​​receives intervention for calculating intervention effects from the user, the presentation unit 34 prohibits receiving intervention from the user for values ​​including the variable X having a number of pieces of data less than a predetermined number.

[0149] Therefore, it is possible to prevent a low-reliability calculation result showing an intervention effect because it is possible to prevent the user from intervening in data having a value of “1” that is an insufficient number of pieces of data.

[0150] Note that Fig.15 In the example, intervention is prohibited by superimposing a display of "not typeable" on the data of the value "1". However, the display mode can be arbitrary. For example, the data of the value for which intervention in the value is prohibited can be displayed lightly or can be displayed transparently. Alternatively, the data can be set not to change even if the user performs an operation to intervene without changing the display mode. In addition, the presentation unit 34 can also display the reason for "not typeable" on the screen. For example, for the value "1", the presentation unit 34 can display that the data cannot be changed due to insufficient number of data items, or that the data cannot be changed due to low reliability of the calculation result of the intervention effect caused by insufficient number of data items.

[0151] <(5) Collecting data again>

[0152] Next, in Fig.16A specific example of the proposed measure “collect data again” is shown in . In a case where the user selects the proposed measure “collect data again”, the presentation unit 34 presents to the user to collect data again while presenting information indicating the variables having insufficient data.

[0153] Specifically, Fig.16 As shown, the presentation unit 34 presents to the user the focused collection of “the combination data of address = “Kanto” and satisfaction level with service A = “4: slightly dissatisfied””, which is a combination of variables having insufficient data.

[0154] Note that in the case of presenting the recollected data, the presenting unit 34 is not limited to the case of performing text display of the variable having insufficient data, and may also present the data in a manner such as Figure 8 The method shown displays information in a table (visual information).

[0155] <<4. Flowchart of information processing>>

[0156] Next, we will refer to Fig.17 An information processing procedure performed by the information processing apparatus 1 according to the present embodiment is described. Fig.17 is a flowchart showing the information processing procedure performed by the information processing device 1 according to the present embodiment. Fig.17 It is assumed that the causal information 41 is generated in advance.

[0157] like Fig.17 As shown, first, the receiving unit 31 receives the user's selection of the intervention variable and the target variable from the variables in the causal information 41 (step S101).

[0158] Subsequently, the extraction unit 32 extracts variables (eg, confounding variables) required for intervention effect calculation from the variables in the causal information 41 (step S102 ).

[0159] Next, the determination unit 33 determines whether there is missing data in the variables extracted by the extraction unit 32 (step S103 ).

[0160] When the determination unit 33 determines that there is missing data in the variable (step S103: Yes), the presentation unit 34 presents information including the following options (1) to (3) to the user, and the reception unit 31 receives a selection of any one of the options from the user (step S104).

[0161] (1) Excluding data with missing values

[0162] (2) Filling in missing values

[0163] (3) Treat missing values ​​as categorical values

[0164] Subsequently, the determination unit 33 performs processing corresponding to the option selected from (1) to (3) on the variables having missing values, aggregating data for each combination of variables required for intervention effect calculation (step S105 ).

[0165] Then, the determination unit 33 determines whether the number of pieces of aggregated data is equal to or greater than a threshold value (step S106 ). That is, the determination unit 33 determines the insufficiency of data of variables required for intervention effect calculation.

[0166] When the number of pieces of aggregated data is equal to or greater than the threshold, that is, when the data of the variables required for intervention effect calculation is sufficient (step S106 : Yes), the presentation unit 34 presents to the user that intervention effect calculation is possible (step S107 ).

[0167] Subsequently, the receiving unit 31 receives a change to the distribution of the intervention variable from the user, and the intervention calculation unit 36 ​​performs intervention effect calculation in the target variable based on the intervention content received from the user (step S108 ).

[0168] Subsequently, the presentation unit 34 presents the change in the distribution of the target variable based on the calculation result of the intervention calculation unit 36 ​​to the user (step S109 ), and ends the processing.

[0169] On the other hand, in a case where there is no missing data in the variables extracted in step S103 (step S103 : No), the control unit 3 advances the process to step S105 .

[0170] In addition, when the number of aggregated data is less than the threshold, that is, when the data of the variables required for intervention effect calculation in step S106 is insufficient (step S106: No), the presentation unit 34 presents to the user that intervention effect calculation is not feasible (step S110).

[0171] Furthermore, the presentation unit 34 displays the list of proposed measures together with the recommendation information indicating the degree of recommendation (step S111). Subsequently, the presentation unit 34 receives selection of one proposed measure from the user and displays information related to the received proposed measure (step S112).

[0172] Subsequently, the proposed measure execution unit 35 determines whether the user performs an operation that permits execution of the proposed measure (step S113 ), and ends the processing if an operation that does not permit execution of the proposed measure is performed (step S113 : No).

[0173] On the other hand, in the case where the user performs an operation that permits execution of the proposed measure (step S113 : Yes), the proposed measure execution unit 35 executes the proposed measure (step S114 ) and advances the process to step S105 .

[0174] <<5. Hardware Configuration Example>>

[0175] Next, we will refer to Fig.18 An example of the hardware configuration of the information processing apparatus 1 or the like according to the present embodiment is described. Fig.18 : is a block diagram showing an example of the hardware configuration of the information processing apparatus 1 according to the present embodiment.

[0176] like Fig.18 As shown, the information processing device 1 includes a central processing unit (CPU) 901, a read-only memory (ROM) 902, a random access memory (RAM) 903, a host bus 905, a bridge 907, an external bus 906, an interface 908, an input device 911, an output device 912, a storage device 913, a drive 914, a connection port 915, and a communication device 916. The information processing device 20 may include an electronic circuit instead of the CPU 901 or in addition to the CPU 901 and a processing circuit such as a DSP or an ASIC.

[0177] The CPU 901 functions as an operation processing device and a control device, and controls the overall operation of the information processing device 20 according to various programs. In addition, the CPU 901 may be a microprocessor. The ROM 902 stores programs, operation parameters, and the like used by the CPU 901. The RAM 903 temporarily stores programs used in the execution of the CPU 901, parameters that change appropriately in the execution, and the like. For example, the CPU 901 may perform the functions of the receiving unit 31, the extracting unit 32, the determining unit 33, the presenting unit 34, the proposed measure executing unit 35, and the intervention calculating unit 36.

[0178] The CPU 901, the ROM 902, and the RAM 903 are connected to each other via a host bus 905 including a CPU bus, etc. The host bus 905 is connected to an external bus 906 such as a peripheral component interconnect / interface (PCI) bus via a bridge 907. Note that the host bus 905, the bridge 907, and the external bus 906 do not necessarily have to be configured in a separate manner, and the functions thereof may be mounted on one bus.

[0179] The input device 911 is a device to which a user inputs information, and is, for example, a mouse, a keyboard, a touch panel, a button, a microphone, a switch, a joystick, etc. Alternatively, the input device 911 may be a remote control device using infrared or other radio waves, or may be an external connection device corresponding to the operation of the information processing device 1, such as a cellular phone or a PDA. In addition, the input device 911 may include, for example, an input control circuit, etc., which generates an input signal based on information input by the user using the above-mentioned input means.

[0180] The output device 912 is a device capable of visually or audibly notifying the user of information. For example, the output device 912 may be a display device such as a cathode ray tube (CRT) display device, a liquid crystal display device, a plasma display device, an electroluminescent (EL) display device, a laser projector, a light emitting diode (LED) projector or a lamp, or may be a sound output device such as a speaker or headphones.

[0181] The output device 912 may, for example, output the results obtained by various processes performed by the information processing device 1. Specifically, the output device 912 may visually display the results obtained by various processes of the information processing device 1 in various forms (e.g., text, images, tables, or graphs). Alternatively, the output device 912 may convert an audio signal such as audio data or sound data into an analog signal and audibly output the analog signal. For example, the input device 911 and the output device 912 may perform the function of an interface.

[0182] The storage device 913 is a device for data storage, and forms an example of the storage unit 4 as the information processing device 1. The storage device 913 can be implemented, for example, by a magnetic storage device such as a hard disk drive (HDD), a semiconductor storage device, an optical storage device, a magneto-optical storage device, or the like. For example, the storage device 913 may include a storage medium, a recording device that records data into the storage medium, a reading device that reads data from the storage medium, a deletion device that deletes data recorded in the storage medium, and the like. The storage device 913 can store a program executed by the CPU 901, various data, various data acquired from the outside, and the like. For example, the storage device 913 can perform a function of storing the causal information 41 and the customer information 42.

[0183] The drive 914 is a reader / writer for a storage medium, and is built in or externally connected to the information processing apparatus 1. The drive 914 reads information recorded in a mounted removable storage medium such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, and outputs it to the RAM 903. Furthermore, the drive 914 can write information to the removable storage medium.

[0184] The connection port 915 is an interface connected to an external device. The connection port 915 is a connection port capable of transmitting data to an external device, and may be, for example, a universal serial bus (USB).

[0185] The communication device 916 is, for example, an interface formed by a communication device for connecting to the network 920, etc. The communication device 916 may be, for example, a communication card for a wired or wireless local area network (LAN), long term evolution (LTE), Bluetooth (registered trademark), or wireless USB (WUSB), etc. In addition, the communication device 916 may be a router for optical communication, a router for asymmetric digital subscriber line (ADSL), a modem for various communications, etc. For example, based on a predetermined protocol such as TCP / IP, the communication device 916 may send / receive signals to / from the Internet or other communication devices, etc.

[0186] Note that the network 40 is a wired or wireless transmission path for information. For example, the network 40 may include a public network such as the Internet, a telephone network, or a satellite communication network, various local area networks (LANs) including Ethernet (registered trademark), a wide area network (WAN), etc. In addition, the network 920 may include a private network such as an Internet Protocol Virtual Private Network (IP-VPN).

[0187] In addition, a computer program that causes hardware such as CPU, ROM, RAM, etc. built into the information processing device 1 to execute functions equivalent to those of the above-described information processing device 1 according to the present embodiment may be created. In addition, a storage medium storing the computer program may be provided.

[0188] In addition, in the multiple processes described in the above embodiments, all or part of the processes described as being automatically performed may be performed manually, or all or part of the processes described as being manually performed may be automatically performed by known methods. In addition, unless otherwise specified, the processing procedures, specific names, and information including various data and parameters shown in the above documents or drawings may be changed arbitrarily. For example, the various information shown in the various figures is not limited to the information shown.

[0189] Moreover, each component of each of the illustrated devices is a functional concept and does not need to be physically configured in the manner shown in the drawings. That is, the specific form of distribution / integration of each device is not limited to that shown in the drawings, and all or part of it can be functionally or physically distributed / integrated in any unit according to various loads and usage conditions.

[0190] In addition, the above-described embodiments can be arbitrarily combined in areas where the processing contents do not contradict each other. In addition, the order of the steps shown in the flowcharts and sequence diagrams of the above-described embodiments can be appropriately changed.

[0191] <<6. Conclusion>>

[0192] As described above, according to an embodiment of the present disclosure, the information processing device 1 includes a determination unit 33 and a presentation unit 34. In the case where calculation is performed on an intervention effect produced in a target variable by intervention of any variable among a plurality of variables based on the causal information 41 indicating a causal relationship between a plurality of variables, the determination unit 33 determines whether data of the variable required for the calculation is insufficient. The presentation unit 34 presents the determination result information based on the determination unit 33 to the user.

[0193] As a result, it is possible to prevent low-precision calculation results from being presented without the user's intention, and it is also possible to inform the user that the current number of data pieces is insufficient. Therefore, appropriate information can be presented to the user according to the data status for performing intervention effect calculation.

[0194] Furthermore, in the case where the determination unit 33 determines that the data of the variables required for the calculation are insufficient, the presentation unit 34 presents information indicating that the calculation is not feasible to the user.

[0195] Therefore, it is possible to prevent low-precision calculation results from being presented.

[0196] Furthermore, in the case where the determination unit 33 determines that the data of the variables required for the calculation are insufficient, the presentation unit 34 presents information indicating the variables having insufficient data to the user.

[0197] Therefore, users can grasp the data with insufficient variables.

[0198] In addition, in the case where the data of a combination of a plurality of variables is insufficient, the presentation unit 34 visualizes information indicating the variables having insufficient data and presents it to the user.

[0199] Thus, the user can intuitively understand that the number of data items in the combination of multiple variables is insufficient for the combination.

[0200] Furthermore, in the case where the determination unit 33 determines that the data of the variables required for the calculation are insufficient, the presentation unit 34 presents to the user proposed measures to satisfy the data of the variables required for the calculation.

[0201] Therefore, it can be presented to the user that the data shortage can be resolved even with the current number of data items.

[0202] Furthermore, as a suggestion measure, the presentation unit 34 presents the user with binding the data of the value having insufficient data with the data of another value among the plurality of values ​​included in the variable.

[0203] Therefore, insufficient data can be solved because the possible values ​​of the variable can be bound and the number of data entries can be added up.

[0204] In addition, as a suggestion measure, the presentation unit 34 presents to the user a variable that has a small influence on the intervention effect among the variables required for the calculation and excludes them from the calculation.

[0205] Therefore, when the influence of a variable with insufficient data is small, the variable can be removed from the intervention effect calculation. Insufficient data of the variable required for the intervention effect calculation can be solved.

[0206] Furthermore, as a suggestion measure, the presentation unit 34 presents the direction of the cause and effect of changing between variables required for calculation to the user.

[0207] Thus, by changing the variables required for intervention effect calculation, the variables having insufficient data become unnecessary for intervention effect calculation, and thus insufficient data of the variables required for intervention effect calculation can be eliminated.

[0208] Furthermore, as a proposed measure, the presentation unit 34 presents to the user to continue the calculation while presenting a decrease in the accuracy of the calculation.

[0209] Therefore, for example, a user who does not want to change the content of data can perform intervention effect calculation.

[0210] Furthermore, as a proposed measure, the presentation unit 34 presents to the user to collect data again while presenting information indicating the variables having insufficient data.

[0211] Therefore, when data is collected again, the user can be presented with which data may be insufficient.

[0212] Furthermore, as a proposed measure, the presentation unit 34 presents to the user a change of the variable to be intervened without changing the target variable.

[0213] Thus, by changing the variables required for intervention effect calculation, the variables having insufficient data become unnecessary for intervention effect calculation, and thus insufficient data of the variables required for intervention effect calculation can be eliminated.

[0214] Furthermore, the presentation unit 34 presents supplementary information related to the proposed measure together with the proposed measure to the user.

[0215] Therefore, the user can understand the content of the proposed measures more deeply.

[0216] Furthermore, in the case of presenting a plurality of proposed measures, the presenting unit 34 presents recommendation information based on the user's proficiency level for each proposed measure.

[0217] Therefore, appropriate proposed measures can be recommended according to the user's proficiency.

[0218] In addition, the information processing device 1 according to the present embodiment further includes an intervention calculation unit 36. In the case where the determination unit 33 determines that the data of the variable required for the calculation is sufficient, the intervention calculation unit 36 ​​calculates the intervention effect of the target variable based on the intervention content received from the user. The presentation unit 34 presents the intervention information based on the calculation result of the intervention calculation unit 36 ​​to the user.

[0219] Therefore, intervention information is not presented if there is insufficient data for a variable, and the calculated results of the intervention effect can be presented to the user only if there is sufficient data.

[0220] Furthermore, in a case where the intervention calculation unit 36 ​​receives intervention for calculating an intervention effect from the user, the presentation unit 34 prohibits receiving intervention from the user for a value having a smaller number of pieces of data than a predetermined number among a plurality of values ​​included in the variable.

[0221] This prevents the user from intervening in data of a value with an insufficient number of data items, thereby preventing the display of a low-reliability calculation result of the intervention effect.

[0222] Furthermore, the presentation unit 34 presents intervention information including the intervention effect and the reliability of the intervention effect to the user.

[0223] Thus, it is possible to indicate how much the user can trust the effect of the intervention as a result of the calculation.

[0224] In addition, the presentation unit 34 presents intervention information to the user, the intervention information including data of the intervention variable and the target variable before and after the intervention.

[0225] This allows the user to easily grasp the status of the intervention variable and the target variable before and after the intervention.

[0226] Although the embodiments of the present disclosure have been described above, the technical scope of the present disclosure is not limited to the above embodiments, and various modifications can be made within the spirit and scope of the present disclosure. Moreover, components of different embodiments and modified examples can be arbitrarily combined.

[0227] Furthermore, the effects in each embodiment described in this specification are merely examples and not limitations, and there may be different effects.

[0228] Note that the present technology may also have the following configurations.

[0229] (1) An information processing device comprising:

[0230] a determination unit that determines whether data of a variable required for calculation is insufficient in a case where calculation of an intervention effect in a target variable produced by an intervention in any one of the plurality of variables is performed based on causal information indicating a causal relationship between the plurality of variables; and

[0231] A presenting unit presents information based on the determination result of the determining unit to a user.

[0232] (2) The information processing device according to (1), wherein

[0233] In a case where the determination unit determines that the data of the variables required for the calculation are insufficient, the presentation unit presents information indicating that the calculation cannot be performed to the user.

[0234] (3) The information processing device according to (1) or (2), wherein

[0235] In a case where the determination unit determines that the data of the variable required for the calculation is insufficient, the presentation unit presents information indicating the variable having the insufficient data to the user.

[0236] (4) The information processing device according to (3), wherein

[0237] In the case where there is insufficient data for a combination of a plurality of variables, the presentation unit visualizes information indicating the variables having insufficient data and presents it to the user.

[0238] (5) The information processing device according to any one of (1) to (4), wherein

[0239] In a case where the determination unit determines that the data of the variables required for the calculation are insufficient, the presentation unit presents to the user proposed measures to satisfy the data of the variables required for the calculation.

[0240] (6) The information processing device according to (5), wherein

[0241] The presenting unit presents, to the user, a proposed measure of binding data of a value having insufficient data and data of another value among a plurality of values ​​included in the variable.

[0242] (7) The information processing device according to (5) or (6), wherein

[0243] The presenting unit presents to the user, as a suggested measure, a variable that has a small influence on the intervention effect among the variables required for the calculation is excluded from the calculation.

[0244] (8) The information processing device according to any one of (5) to (7), wherein

[0245] The presentation unit presents the direction of changing the cause and effect between the variables required for the calculation to the user as a proposed measure.

[0246] (9) The information processing device according to any one of (5) to (8), wherein

[0247] The presenting unit presents to the user that the calculation is continued while presenting a decrease in the accuracy of the calculation as a proposed measure.

[0248] (10) The information processing device according to any one of (5) to (9), wherein

[0249] The presenting unit presents to the user, as a suggested measure, to collect data again while presenting information indicating the variable having insufficient data.

[0250] (11) The information processing device according to any one of (5) to (10), wherein

[0251] The presenting unit presents, to the user, changes to the variable to be intervened but not to the target variable as proposed measures.

[0252] (12) The information processing device according to any one of (5) to (11), wherein

[0253] The presenting unit presents the supplementary information related to the proposed measure to the user together with the proposed measure.

[0254] (13) The information processing device according to any one of (5) to (12), wherein

[0255] In the case where a plurality of proposed measures are presented, the presentation unit presents recommendation information for each proposed measure, the recommendation information being based on the user's proficiency level.

[0256] (14) The information processing device according to any one of (1) to (13), further comprising an intervention calculation unit, which performs calculation of the intervention effect in the target variable based on the intervention content received from the user when the determination unit determines that the data of the variable required for the calculation is sufficient, wherein

[0257] The presentation unit presents intervention information based on the calculation result of the intervention calculation unit to a user.

[0258] (15) The information processing device according to (14), wherein

[0259] In a case where the intervention calculation unit receives an intervention for calculating an intervention effect from a user, the presentation unit prohibits receiving the intervention for a value having a number of data pieces less than a predetermined number among a plurality of values ​​included in the variable.

[0260] (16) The information processing device according to (14) or (15), wherein

[0261] The presentation unit presents intervention information including the intervention effect and the reliability of the intervention effect to the user.

[0262] (17) The information processing device according to any one of (14) to (16), wherein

[0263] The presenting unit presents intervention information to a user, where the intervention information includes data of an intervention variable and a target variable before and after the intervention.

[0264] (18) An information processing method comprising:

[0265] a determining step of, in a case where an intervention effect in a target variable due to intervention on any one of a plurality of variables is calculated based on causal information indicating a causal relationship between a plurality of variables, determining whether data of a variable required for the calculation is insufficient; and

[0266] The presenting step presents information based on the determination result of the determining step to the user.

[0267] (19) An information processing program which, when read by a computer, causes the computer to function as

[0268] a determination unit that determines whether data of a variable required for calculation is insufficient when calculating an intervention effect in a target variable due to intervention on any one of the plurality of variables based on causal information indicating a causal relationship between the plurality of variables, and

[0269] A presenting unit presents information based on the determination result of the determining unit to a user.

[0270] Reference numerals list

[0271] 1 Information processing equipment

[0272] 2 Communication unit

[0273] 3 Control unit

[0274] 4 Storage Units

[0275] 11 User Terminal

[0276] 31 Receiving unit

[0277] 32 Extraction Units

[0278] 33 Determine the unit

[0279] 34 Presentation Units

[0280] 35 Proposed measures implementation unit

[0281] 36 Intervention Computing Unit

[0282] 41 Causal Information

[0283] 42 Customer Information

Claims

1. An information processing device for performing data analysis in the field of providing products or services to customers, comprising: a determination unit that determines whether data of a variable required for calculating an intervention effect in a target variable by intervening on any one of a plurality of variables is insufficient, in a case where the intervention effect is calculated based on causal information indicating a causal relationship between the plurality of variables, wherein the causal information is generated based on attribute data of a customer and questionnaire data of the customer; and a presenting unit, which presents information to a user based on the determination result of the determining unit, The determining unit performs aggregation processing of the aggregated data for each combination of the intervention variable, the target variable, and the variable required for the calculation, and determines for each combination whether the number of data items is equal to or greater than a predetermined threshold, and determines that the data of the variable required for the calculation is sufficient when the number of data items in all combinations is equal to or greater than the predetermined threshold, and determines that the data of the variable required for the calculation is insufficient when there is a combination of variables whose number of data items is less than the predetermined threshold in all combinations. In response to a correlation between the intervention variable and one of the multiple variables being equal to or greater than another predetermined threshold, the presentation unit presents to the user proposed measures for changing the intervention variable to the variable without changing the target variable as data for the variable required to satisfy the calculation.

2. The information processing device according to claim 1, wherein: In a case where the determination unit determines that the data of the variable required for the calculation is insufficient, the presentation unit presents information indicating that the calculation is not feasible to the user.

3. The information processing device according to claim 1, wherein: In a case where the determination unit determines that data of the variable required for the calculation is insufficient, the presentation unit presents information indicating the variable having insufficient data to the user.

4. The information processing device according to claim 3, wherein: In the case where there is insufficient data for a combination of a plurality of variables, the presentation unit visualizes information indicating the variables having insufficient data and presents it to the user.

5. The information processing device according to claim 1, wherein: In a case where the determination unit determines that the data of the variable required for the calculation is insufficient, the presentation unit presents to the user proposed measures for satisfying the data of the variable required for the calculation.

6. The information processing device according to claim 5, wherein: The presentation unit presents, to the user, binding of data having a value with insufficient data and data of another value with respect to a plurality of values ​​included in the variable as the proposed measure.

7. The information processing device according to claim 5, wherein: The presentation unit presents, to the user, as the proposed measure, a variable whose influence on the intervention effect is less than a predetermined threshold value among the variables required for the calculation, which is excluded from the calculation.

8. The information processing device according to claim 5, wherein: The presentation unit presents, to the user, a direction of changing the cause and effect between the variables required for the calculation as the proposed measure.

9. The information processing device according to claim 5, wherein: The presentation unit presents to the user that the calculation is continued while presenting a decrease in the accuracy of the calculation as the proposed measure.

10. The information processing device according to claim 5, wherein: The presenting unit presents, to the user, collecting data again while presenting information indicating a variable having insufficient data as the proposed measure.

11. The information processing device according to claim 5, wherein: The presenting unit presents supplementary information related to the proposed measure to the user together with the proposed measure.

12. The information processing device according to claim 5, wherein: In the case where a plurality of the proposed measures are presented, the presentation unit presents recommendation information for each of the proposed measures, the recommendation information being based on a user's proficiency level.

13. The information processing device according to claim 1, further comprising an intervention calculation unit, which performs calculation of the intervention effect of the target variable based on the intervention content received from the user when the determination unit determines that the data of the variable required for the calculation is sufficient, wherein The presentation unit presents intervention information based on a calculation result of the intervention calculation unit to the user.

14. The information processing device according to claim 13, wherein: In a case where the intervention calculation unit receives an intervention from the user to calculate the intervention effect, the presentation unit prohibits receiving an intervention on a value having a number of data pieces less than a predetermined number among a plurality of values ​​included in a variable.

15. The information processing device according to claim 13, wherein: The presentation unit presents the intervention information including the intervention effect and the reliability of the intervention effect to the user.

16. The information processing device according to claim 13, wherein: The presentation unit presents the intervention information including the intervention variable and the data of the target variable before and after the intervention to a user.

17. An information processing method for performing data analysis in the field of providing products or services to customers, comprising: a determining step of determining whether data of a variable required for calculating an intervention effect in a target variable produced by intervening in any one of a plurality of variables based on causal information indicating a causal relationship between a plurality of variables, wherein the causal information is generated based on attribute data of a customer and questionnaire data of the customer, is insufficient; and a presenting step of presenting information based on the determination result of the determining step to a user, The determining step includes performing aggregation processing of aggregated data for each combination of the intervention variable, the target variable, and the variable required for the calculation, and for each combination, determining whether the number of data items is equal to or greater than a predetermined threshold, and determining that the data of the variable required for the calculation is sufficient when the number of data items in all combinations is equal to or greater than the predetermined threshold, and determining that the data of the variable required for the calculation is insufficient when there is a combination of variables whose number of data items is less than the predetermined threshold in all combinations. The presenting step also includes presenting to the user, in response to a correlation between the intervening variable and a variable among the plurality of variables being equal to or greater than another predetermined threshold, proposed measures of changing the intervening variable to the variable without changing the target variable as data for satisfying the variable required for the calculation.

18. A computer-readable storage medium storing an information processing program for performing data analysis in providing products or services to customers, which is read by a computer to cause the computer to function as: a determination unit that determines whether data of a variable required for calculation is insufficient in a case where an intervention effect produced in a target variable by intervening in any one of a plurality of variables is calculated based on causal information indicating a causal relationship between a plurality of variables, wherein the causal information is generated based on attribute data of a customer and questionnaire data of the customer, and a presentation unit that presents information based on a determination result of the determination unit to a user, in, The determination unit performs aggregation processing of the aggregated data for each combination of the intervention variable, the target variable, and the variable required for the calculation, and for each combination, determines whether the number of data items is equal to or greater than a predetermined threshold, and determines that the data of the variable required for the calculation is sufficient when the number of data items in all combinations is equal to or greater than the predetermined threshold, and determines that the data of the variable required for the calculation is insufficient when there is a combination of variables whose number of data items is less than the predetermined threshold in all combinations, In response to a correlation between the intervention variable and one of the multiple variables being equal to or greater than another predetermined threshold, the presentation unit presents to the user proposed measures for changing the intervention variable to the variable without changing the target variable as data for the variable required to satisfy the calculation.

Images