Inference device

Abstract

An inference device according to the present invention comprises a first calculation unit, a second calculation unit, and an inference unit. The first calculation unit calculates a predicted value for an objective variable for each of a plurality of areas on the basis of a plurality of types of explanatory variables that include a first explanatory variable. The second calculation unit calculates indicators that indicate the strength of the correlation between the difference between the predicted value and a measured value for the objective variable and each of a plurality of second explanatory variables obtained by dividing the first explanatory variable. The inference unit compares the indicators calculated by the second calculation unit for each of the plurality of second explanatory variables and thereby infers second explanatory variables that contribute to an increase in the objective variable.

Description

estimation device

[0001] The present disclosure relates to an estimation device.

[0002] Patent Literature 1 discloses an invention of a system that determines a display priority for explanatory variables used to calculate the value of a dependent variable and displays the explanatory variables. In this system, the display priority of each explanatory variable used to calculate the value of the dependent variable using an estimation formula is determined based on the magnitude of the influence of the explanatory variable on the difference between the actual measured value of the dependent variable and the value calculated according to the estimation formula. Note that the dependent variable is a variable that is desired to be estimated, such as future sales, and the explanatory variable is a variable that provides a clue to estimating the value of the dependent variable, such as sales over the past two weeks. The estimation formula is an equation that represents a regularity found between the dependent variable and the explanatory variables.

[0003] International Publication No. 2016 / 063446

[0004] In recent years, various payment services have been offered to smartphone users, allowing them to make payments at stores and award points based on the payment amount. Increasing the number of active users is crucial to the widespread adoption of these payment services. An active user is a member who has registered to use a payment service and has used the service within a specified period (e.g., the most recent month). To increase the number of active users, payment service providers conduct sales activities to increase the number of affiliated stores that accept their payment services. Due to budget constraints, it is desirable to focus sales activities on industries that will significantly contribute to an increase in active users. However, there has been no technology available to date that can identify industries that significantly contribute to an increase in active users.

[0005] An estimation device according to a preferred aspect of the present disclosure includes a first calculation unit, a second calculation unit, and an estimation unit. The first calculation unit calculates a predicted value of a dependent variable for each of a plurality of areas based on a plurality of explanatory variables including a first explanatory variable. The second calculation unit calculates an index indicating the strength of correlation between the difference between the actual value and the predicted value of the dependent variable and each of a plurality of second explanatory variables obtained by subdividing the first explanatory variable. The estimation unit estimates a second explanatory variable that contributes to an increase in the dependent variable by comparing the index calculated by the second calculation unit for each of the plurality of second explanatory variables.

[0006] According to the estimation device of the present disclosure, when a first explanatory variable included in a plurality of explanatory variables used to calculate a predicted value of a dependent variable is subdivided into a plurality of second explanatory variables, it is possible to estimate a second explanatory variable among the plurality of second explanatory variables that contributes to an increase in the dependent variable.

[0007] FIG. 1 is a block diagram showing an example configuration of an estimation device 10A according to a first embodiment of the present disclosure. FIG. 2 is a flowchart showing a processing flow in an estimation method executed by a processing device 130 of the estimation device 10A according to a program PR1. FIG. 3 is a block diagram showing an example configuration of an estimation device 10B according to a second embodiment of the present disclosure. FIG. 4 is a diagram for explaining estimation by an estimation unit 130c of the estimation device 10B. FIG. 5 is a flowchart showing a processing flow in an estimation method executed by a processing device 130 of the estimation device 10B according to a program PR2.

[0008] 1 is a block diagram showing an example configuration of an estimation device 10A according to a first embodiment of the present disclosure. The estimation device 10A has a function of calculating, by machine learning or the like, a predicted value of an active rate for each area where payment services such as points awarded in conjunction with payment at a store are provided, based on a plurality of explanatory variables such as the member ratio for the area and the density of affiliated stores.

[0009] The payment service in this embodiment is a service provided to users of mobile devices such as smartphones. The payment service is provided by a business operator that provides communication services such as voice calls to the mobile devices. In this embodiment, an area refers to an administrative unit such as a city, ward, town, or village. One city, one town, one ward, and one village each constitute one area. Note that a government-designated city may be treated as one area on a city-by-city basis.

[0010] The active rate of a payment service is the ratio of the number of active users AU to the number of members of the payment service in an area TU, i.e., AU / UT. An active user is a member who has registered to use a payment service and has used the payment service within a specified period (e.g., the most recent month). Note that use of a payment service refers to receiving points or using points, such as applying earned points to a payment amount. The active rate is calculated on a monthly and area basis. The active rate is an example of a target variable in this disclosure, i.e., a variable whose value is to be estimated.

[0011] The predicted value of the active rate is calculated based on multiple explanatory variables. The multiple explanatory variables include, for example, the membership rate, the affiliated store density, and the reach means consent rate. The membership rate is the ratio of the number of payment service members (TU) in an area to the area's population (TP), i.e., TU / TP. The affiliated store density is the ratio of the number of stores (TS) of all industries that offer payment services in the area to the area's area (SA), i.e., TS / SA. The affiliated store density can be said to be the number of affiliated stores across all industries per area's population. The reach means consent rate is the ratio of the number of users (TA) in the area who consent to information distribution from payment service providers to the area's population (TP), i.e., TA / TP. Payment service information distribution can be, for example, sending direct mail, email distribution, and push notifications to apps for using the payment service.

[0012] As mentioned above, to widely popularize payment services, it is important to increase the number of active users, i.e., to improve the active rate. To improve the active rate, it is important to increase the number of affiliated stores that offer payment services, in addition to encouraging members to use the service. This is because the more affiliated stores that offer payment services within an area, the greater the convenience in that area and the greater the expected use of the payment service. Payment service providers conduct sales activities to increase the number of affiliated stores that offer their payment services, but there are budgetary constraints for these sales activities. Therefore, it is desirable to focus sales activities on stores in industries that strongly contribute to improving the active rate. The estimation device 10A is a device that makes it possible to narrow down the industries that strongly contribute to improving the active rate.

[0013] As shown in FIG. 1 , the estimation device 10A includes a communication device 110, a storage device 120, a processing device 130, and a bus 140. The communication device 110, the storage device 120, and the processing device 130 are connected to each other by the bus 140 that mediates data exchange. The bus 140 may be configured using a single bus, or may be configured using different buses between each element. Although not shown in detail in FIG. 1 , the estimation device 10A includes an operation unit that accepts input operations by workers who narrow down the industries, and a display unit that displays various information.

[0014] The communication device 110 is hardware (transmission / reception device) for communicating with other devices. Other devices that communicate with the estimation device 10A are connected to the communication device 110 via wired or wireless connections. A specific example of another device connected to the communication device 110 is a payment server that provides payment services. The payment server stores information about payment services by area and month. The information about payment services includes, for example, the population within the area, the area size, the number of members within the area, the number of times each member has used the payment service, the most recent date and time of use, the number of affiliated stores across all industries, the number of affiliated stores by industry, and the reach means acceptance rate. The communication device 110 acquires information about payment services by communicating with the payment server.

[0015] The storage device 120 is a recording medium readable by the processing device 130. The storage device 120 may be configured with at least one of, for example, a read-only memory (ROM), an erasable programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), a random access memory (RAM), etc. The storage device 120 stores a prediction model MDL and a program PR1 in advance.

[0016] The prediction model MDL is a machine learning model that outputs a predicted value of the active rate in response to input of multiple explanatory variables. For example, when a predicted value y1 of the active rate is formulated as y1 = a1 x X1 + a2 x X2 + a3 x X3 ... using multiple explanatory variables X1, X2, X3 ..., multiple training data are prepared. Each of the multiple training data indicates multiple explanatory variables and an actual measured value of the active rate. The multiple training data are used to determine parameters a1, a2, a3 ... so that the difference between the predicted value y1 and the actual measured value of the active rate, i.e., the value obtained by calculating "predicted value y1 - actual measured value," is approximately zero. The prediction model MDL is generated by determining the parameters a1, a2, a3 ... using the multiple training data. The parameters a1, a2, a3 ... are weights that represent the magnitude of the contribution of the explanatory variables X1, X2, X3 ... to the predicted value y1. In this embodiment, the multiple explanatory variables include the member ratio, the affiliated store density across all industries, and the reach means consent rate. However, the multiple explanatory variables in this embodiment do not include the affiliated store density for each industry. As the machine learning algorithm for generating the prediction model MDL, an existing algorithm can be appropriately adopted.

[0017] The processing device 130 includes one or more central processing units (CPUs). The processing device 130 reads the program PR1 from the storage device 120 in response to an instruction to execute the program PR1 via an input operation on an operation unit. The processing device 130 executes the read program PR1. By executing the program PR1, the processing device 130 functions as the first calculation unit 130a, the second calculation unit 130b, and the estimation unit 130c shown in FIG. 1 . In other words, the first calculation unit 130a, the second calculation unit 130b, and the estimation unit 130c shown in FIG. 1 are software modules realized by operating a computer such as a CPU in accordance with software such as a program. The functions of the first calculation unit 130a, the second calculation unit 130b, and the estimation unit 130c are as follows:

[0018] The first calculation unit 130a calculates a predicted value of the active rate for each of a plurality of areas using information acquired from the payment server via the communication device 110 and the prediction model MDL. More specifically, the first calculation unit 130a calculates the membership ratio from the population and number of members in the area acquired using the communication device 110. The first calculation unit 130a also calculates the affiliated store density across all industries based on the number of affiliated stores across all industries in the area and the area size acquired using the communication device 110. The first calculation unit 130a then calculates a predicted value of the active rate by inputting the reach means permission rate acquired using the communication device 110, the membership ratio calculated as described above, and the affiliated store density across all industries into the prediction model MDL as explanatory variables. The prediction model MDL is executed by the processing device 130 (e.g., the first calculation unit 130a). The predicted value output from the prediction model MDL in accordance with the explanatory variables input to the prediction model MDL is used as the predicted value of the active rate.

[0019] The second calculation unit 130b calculates an actual active rate for each of multiple areas where the payment service is provided based on information acquired from the payment server by the communication device 110. Specifically, the second calculation unit 130b calculates the actual active rate by dividing the number of members in the area who used the payment service during a predetermined period (e.g., the most recent month) by the total number of members in the area. The second calculation unit 130b also calculates the merchant density for each industry based on the number of affiliated stores for each industry acquired by the communication device 110. The merchant density for a certain industry is the ratio of the number of stores in that industry that offer the payment service in that area to the area's surface area. Since the merchant density for each industry is the total merchant density for all industries, the merchant density for each industry can be considered a value obtained by subdividing the total merchant density for all industries by industry. The total merchant density for all industries is an example of a first explanatory variable in the present disclosure. The affiliated store density for each industry is an example of multiple second explanatory variables obtained by subdividing the first explanatory variable. The second calculation unit 130b calculates the difference between the predicted value calculated by the first calculation unit 130a and the actual measured value of the active rate for each of the multiple areas. Hereinafter, the difference between the predicted value and the actual measured value of the active rate, i.e., the "predicted value - actual value," is referred to as the residual. The second calculation unit 130b then calculates an index indicating the strength of the correlation between the residual and the affiliated store density for each industry, targeting all of the multiple areas where payment services are provided. In this embodiment, the Pearson's correlation coefficient, which indicates a linear relationship, is used as the index indicating the strength of the correlation, but an index indicating a rank correlation or a nonlinear relationship may also be used.

[0020] The estimation unit 130c estimates the industries that will contribute to an increase in the active rate by comparing the indices calculated for each industry by the second calculation unit 130b. For example, the estimation unit 130c estimates K industries in descending order of correlation with the residual, i.e., in descending order of correlation coefficient, as industries that will contribute to an increase in the active rate. Note that K is an integer greater than or equal to 1 and less than or equal to the number of industries. The estimation unit 130c then displays information indicating the industries that are estimated to contribute to an increase in the active rate on the display unit. By visually checking the information displayed on the display unit, the worker can understand the industries that are estimated to contribute to an increase in the active rate.

[0021] It should be noted here that the strong correlation between the residual and the franchisee density for each industry means that the residual cannot be considered to be zero. Conventionally, when generating a prediction model, parameters are adjusted so that the residual becomes zero. If the residual does not become zero, it is common to add new types of explanatory variables to improve the explanatory power of the prediction model. In the present disclosure, if the residual between the predicted value calculated using the prediction model MDL generated so that the residual becomes zero and the actual measured value does not become zero, a first process is executed to search for factors that increased the residual from among multiple second explanatory variables that were not used in calculating the predicted value. By executing the first process, it is possible to estimate industries that will actualize latent demand, i.e., industries that will contribute to an increase in the active rate. This is one feature of the present disclosure.

[0022] The processing unit 130, operating in accordance with the program PR1, executes an estimation method specific to this embodiment. Figure 2 is a flowchart showing the process flow of the estimation method of this embodiment. As shown in Figure 2, this estimation method includes a first calculation process SA110, a second calculation process SA120, and an estimation process SA130. The process contents of each process shown in Figure 2 are as follows.

[0023] In the first calculation process SA110, the processing device 130 functions as a first calculation unit 130a. In the first calculation process SA110, the processing device 130 calculates a predicted value of the activity rate, which is the objective variable, for each of the multiple areas based on multiple types of explanatory variables.

[0024] In the second calculation process SA120, the processing device 130 functions as the second calculation unit 130b. In the second calculation process SA120, the processing device 130 calculates an index indicating the strength of the correlation between the residual and the affiliated store density for each industry across all of the multiple areas where the payment service is provided.

[0025] In estimation process SA130, processing device 130 functions as estimation unit 130c. In estimation process SA130, processing device 130 estimates the franchise store density of industries that will contribute to an increase in the active rate by comparing the indices calculated for each industry in second calculation process SA120. Then, in estimation process SA130, processing device 130 displays information indicating the industries that are estimated to contribute to an increase in the active rate on the display unit. By visually checking the information displayed on the display unit, the worker can understand the industries that are estimated to contribute to an increase in the active rate.

[0026] As described above, according to the estimation device 10A of this embodiment, when the franchisee density across all industries, which is one of the multiple explanatory variables used in calculating the predicted value of the active rate, which is the objective variable, is subdivided into franchisee densities for each industry, it is possible to estimate the franchisee density of industries that contribute to an increase in the active rate among the franchisee densities for each industry, i.e., the industries that contribute to an increase in the active rate. According to this embodiment, sales activities can be narrowed down to stores in industries that are estimated to contribute to an increase in the active rate.

[0027] B: Second Embodiment Fig. 3 is a diagram illustrating a configuration example of an estimation device 10B according to a second embodiment of the present disclosure. Similar to the estimation device 10A, the estimation device 10B has a function of calculating, by machine learning or the like, a predicted value of an active rate for each area where payment services such as points awarded in conjunction with payments at stores are provided, based on a plurality of explanatory variables such as the membership rate for the area and the density of affiliated stores. Also, similar to the estimation device 10A, the estimation device 10B has a function of estimating an industry that will strongly contribute to improving the active rate.

[0028] In FIG. 3, the same components as those in FIG. 1 are denoted by the same reference numerals as those in FIG. 1. As is clear from comparing FIG. 3 with FIG. 1, the hardware configuration of the estimation device 10B is the same as the hardware configuration of the estimation device 10A. That is, the estimation device 10B includes a communication device 110, a storage device 120, and a processing device 130. Although not shown in detail in FIG. 3, the estimation device 10B, like the estimation device 10A, includes an operation unit that accepts input operations from workers who narrow down the industries, and a display unit that displays various information. The estimation device 10B differs from the estimation device 10A in that a program PR2 is stored in the storage device 120 instead of the program PR1.

[0029] The processing device 130 of the estimation device 10B reads the program PR2 from the storage device 120 in response to an instruction to execute the program PR2 via an input operation on the operation unit. The processing device 130 executes the read program PR2. By executing the program PR2, the processing device 130 functions as the first calculation unit 130a, extraction unit 130d, second calculation unit 130e, and estimation unit 130c shown in FIG. 3. In other words, the first calculation unit 130a, extraction unit 130d, second calculation unit 130e, and estimation unit 130c shown in FIG. 3 are software modules realized by operating a computer such as a CPU in accordance with software such as a program. The functions of the extraction unit 130d and second calculation unit 130e, which are differences from the first embodiment, are as follows.

[0030] The extraction unit 130d first calculates an actual active rate for each of a plurality of areas where payment services are provided, based on information acquired from the payment server by the communication device 110. The extraction unit 130d then classifies the plurality of areas into a first group and a second group. The first group is a group in which the difference between the predicted value calculated by the first calculation unit 130a and the actual active rate exceeds a second threshold. The second group is a group different from the first group. The second threshold is a non-negative value. In this embodiment, the second threshold is zero. Furthermore, in this embodiment, the second group includes areas in which the difference between the predicted value calculated by the first calculation unit 130a and the actual active rate is equal to or less than a third threshold. The third threshold is a value equal to or less than the second threshold. In this embodiment, the third threshold is the same value as the second threshold, i.e., zero.

[0031] In this embodiment, the second threshold is zero. Therefore, the first group includes areas where the difference between the predicted value and the actual measured value is positive, i.e., areas where the actual measured value is lower than the predicted value. The fact that the actual measured value is lower than the predicted value means that although a higher active rate is predicted from multiple explanatory variables, the actual active rate is low. For areas where the actual measured value is lower than the predicted value, further improvement in the active rate is expected depending on sales activities, i.e., there is expected to be latent demand. For this reason, hereinafter, areas belonging to the first group will be referred to as "areas with high potential."

[0032] In this embodiment, the third threshold is zero, just like the second threshold. Therefore, the second group includes areas where the difference between the predicted value and the actual measured value is zero or negative, i.e., areas where the actual measured value is equal to or greater than the predicted value. The fact that the actual measured value is equal to or greater than the predicted value means that the actual active rate is equal to or greater than the active rate predicted from multiple explanatory variables. Hereinafter, areas that belong to the second group will be referred to as "areas that are more active than their potential."

[0033] Next, the extraction unit 130d extracts N areas from the areas belonging to the first group in descending order of residual error. N is an arbitrary integer greater than or equal to 1 and less than or equal to the total number of areas belonging to the first group. The N areas extracted from the areas belonging to the first group in descending order of residual error are an example of a first area in the present disclosure. Furthermore, the extraction unit 130d extracts M areas from the areas belonging to the second group in descending order of residual error absolute value. M is an arbitrary integer greater than or equal to 1 and less than or equal to the total number of areas belonging to the second group. The M areas extracted from the areas belonging to the second group in descending order of residual error absolute value are an example of a second area in the present disclosure.

[0034] The second calculation unit 130e calculates an index indicating the strength of the correlation between the member store density and the residual for each industry type for the N+M areas extracted by the extraction unit 130d. In this embodiment, as in the first embodiment, the Pearson's correlation coefficient, which indicates a linear relationship, is used as the index indicating the strength of the correlation. However, an index indicating a rank correlation or a non-linear relationship may also be used as the index indicating the strength of the correlation. The second calculation unit 130b in the first embodiment calculated an index indicating the strength of the correlation between the member store density and the residual for each industry type for all of the multiple areas in which payment services are provided. The second calculation unit 130e in this embodiment differs from the second calculation unit 130b in that it calculates an index indicating the strength of the correlation between the member store density and the residual for each industry type for the first area and the second area. The estimation unit 130c estimates the franchise store density of the industries that will contribute to an increase in the active rate by comparing the indices calculated for each industry by the second calculation unit 130e with each other, and displays information indicating the industries that are estimated to contribute to an increase in the active rate on the display unit, as in the first embodiment. The worker can grasp the industries that are estimated to contribute to an increase in the active rate by visually checking the information displayed on the display unit.

[0035] For example, suppose that franchisee businesses are subdivided into three types: convenience stores, supermarkets, and drug stores. The relationship between the franchisee density across all businesses and the franchisee density for each of convenience stores, supermarkets, and drug stores and the residual is as shown in Figure 4. The circles in Figure 4 correspond to the first area, and the triangles in Figure 4 correspond to the second area. The second calculation unit 130e uses the data marked with circles and triangles in Figure 4 to calculate an index showing the strength of the correlation between the residual and the franchisee density for each business type.

[0036] In this embodiment, if the residual between the predicted value calculated using the prediction model MDL generated to achieve zero residuals and the actual measured value is not zero, a first process is executed to identify the factor that increased the residual from among multiple second explanatory variables that were not used to calculate the predicted value. By executing the first process, it is possible to estimate the industry that will actualize latent demand, i.e., the industry that will contribute to an increase in the active rate. In addition, in this embodiment, for example, if the total number of areas in which payment services are provided is 1,000, a second process is executed to set N and M to values ​​between 5 and 10, which are sufficiently small compared to the total number of areas. By executing the second process, it is possible to reduce the calculation processing load required to estimate the industry that will contribute to an increase in the active rate, compared to the first embodiment.

[0037] The processing device 130, operating in accordance with the program PR2, executes an estimation method specific to this embodiment. FIG. 5 is a flowchart showing the processing flow of the estimation method of this embodiment. As shown in FIG. 5, this estimation method includes a first calculation process SA110, an extraction process SB120, a second calculation process SB130, and an estimation process SA130. As is clear from comparing FIG. 5 with FIG. 2, the estimation method of this embodiment differs from the estimation method of the first embodiment in the following two points. The first difference is that the extraction process SB120 is executed following the first calculation process SA110. The second difference is that the second calculation process SB130 is executed instead of the second calculation process SA120. The processing contents of each step of the extraction process SB120 and the second calculation process SB130, which are differences from the estimation method of the first embodiment, are as follows.

[0038] In extraction process SB120, processing device 130 functions as extraction unit 130d. In extraction process SB120, processing device 130 extracts N first areas and M second areas based on the actual measurement value of the active rate and the predicted value calculated in first calculation process SA110.

[0039] In second calculation process SB130, processing device 130 functions as second calculation unit 130e. In second calculation process SB130, processing device 130 calculates an index indicating the strength of correlation between the residual and the affiliated store density for each industry for the N first areas and M second areas extracted in extraction process SB120. The residual is the difference between the predicted value calculated in first calculation process SA110 and the actual measured value of the active rate.

[0040] Thereafter, in estimation process SA130, processing device 130 functions as estimation unit 130c, and estimates the franchise store density of the industry that will contribute to an increase in the active rate by mutually comparing the indices calculated for each industry in second calculation process SB130. Then, in estimation process SA130, processing device 130 displays information indicating the industry that is estimated to contribute to an increase in the active rate on the display unit.

[0041] As described above, the estimation device 10B of the present embodiment, like the estimation device 10A of the first embodiment, can estimate the types of businesses that contribute to an increase in the active rate, and can narrow down sales activities to stores in the types of businesses that are estimated to contribute to an increase in the active rate. In addition, according to the present embodiment, by setting N and M to values ​​that are sufficiently small compared to the total number of areas where payment services are provided, it is possible to reduce the processing load of calculations required to estimate the types of businesses that contribute to an increase in the active rate, compared to the first embodiment.

[0042] C: Modifications The present disclosure is not limited to the first and second embodiments exemplified above. Specific modifications are as follows. Two or more embodiments arbitrarily selected from the following examples may be combined.

[0043] C-1: Variation 1 In the above embodiments, areas are defined by city, ward, town, or village. However, each prefecture may be treated as an area, or an area may be defined by a regional division by an administrative body. For example, in the case of Tokyo, an area may be defined by each of the 23 wards, the Tama region, and the islands. Furthermore, areas are not limited to administrative or regional divisions such as prefectures or city, ward, town, or village. For example, a grid of 1 km square may be treated as an area. Furthermore, the member ratio, which is one of the multiple explanatory variables in the above embodiments, may be the ratio of the number of payment service members to the number of mobile communication service contracts in the area. Furthermore, the affiliated store density, which is also one of the explanatory variables, may be the number of affiliated stores of all industries relative to the population of the area.

[0044] C-2: Modification 2 In each of the above embodiments, the dependent variable is the active rate of the payment service, the first explanatory variable is the affiliated store density across all industries, and the second explanatory variable is the affiliated store density for each industry. However, the dependent variable, first explanatory variable, and second explanatory variable are not limited to these. As long as the estimation device estimates a predicted value of the dependent variable based on multiple explanatory variables including the first explanatory variable, and the first explanatory variable is subdivided into multiple second explanatory variables, the present disclosure can be applied regardless of the dependent variable, first explanatory variable, and second explanatory variable.

[0045] C-3: Modification 3 In the first embodiment, the program PR1 is stored in the storage device 120 of the estimation device 10A. However, the program PR1 may be manufactured or sold separately. The program PR1 may be provided to the purchaser, for example, by distributing a computer-readable recording medium, such as a flash ROM, to which the program PR1 is written, or by downloading the program PR1 via a telecommunications line. Similarly, the program PR2 in the second embodiment may be manufactured or sold separately.

[0046] C-4: Modification 4 In the first embodiment, the first calculation unit 130a, the second calculation unit 130b, and the estimation unit 130c are all software modules. However, any one, any two, or all of the first calculation unit 130a, the second calculation unit 130b, and the estimation unit 130c may be hardware modules. Examples of hardware modules include a digital signal processor (DSP), an application specific integrated circuit (ASIC), a programmable logic device (PLD), and a field programmable gate array (FPGA). Even if any one, any two, or all of the first calculation unit 130a, the second calculation unit 130b, and the estimation unit 130c are hardware modules, the same effects as those of the first embodiment can be achieved. Similarly, any one, any two, any three, or all of the first calculation unit 130a, extraction unit 130d, second calculation unit 130e, and estimation unit 130c in the second embodiment may be hardware modules.

[0047] D: Others (1) In the above-described embodiment, ROM, RAM, etc. are exemplified as storage device 120, but storage device 120 may also be a flexible disk, a magneto-optical disk (e.g., a compact disk, a digital versatile disk, a Blu-ray (registered trademark) disk), a smart card, a flash memory device (e.g., a card, a stick, a key drive), a CD-ROM (Compact Disc-ROM), a register, a removable disk, a hard disk, a floppy (registered trademark) disk, a magnetic strip, a database, a server, or other suitable storage medium.

[0048] (2) In the above-described embodiments, the described information, signals, etc. may be represented using any of a variety of different technologies. For example, data, instructions, commands, information, signals, bits, symbols, chips, etc. that may be referred to throughout the above description may be represented by voltages, currents, electromagnetic waves, magnetic fields or magnetic particles, optical fields or photons, or any combination thereof.

[0049] (3) In the above-described embodiment, input and output information may be stored in a specific location (for example, a memory) or may be managed using a management table. Input and output information may be overwritten, updated, or added to. Output information may be deleted. Input information may be transmitted to another device.

[0050] (4) In the above-described embodiment, the determination may be made based on a value represented by one bit (0 or 1), a Boolean value (true or false), or a comparison of numerical values ​​(e.g., comparison with a predetermined value).

[0051] (5) The order of the process procedures, sequences, flowcharts, etc. illustrated in the above-described embodiments may be rearranged unless it is consistent. For example, the methods described in this disclosure present elements of various steps using an example order, and are not limited to the particular order presented.

[0052] (6) Each function illustrated in Figure 1 or 3 is realized by any combination of hardware and / or software. Furthermore, the method of realizing each functional block is not particularly limited. That is, each functional block may be realized using a single device that is physically or logically coupled, or may be realized using two or more physically or logically separated devices that are directly or indirectly connected (e.g., wired, wireless, etc.) and these multiple devices. A functional block may be realized by combining software with the single device or the multiple devices.

[0053] (7) The programs exemplified in the above-described embodiments should be broadly construed to mean instructions, instruction sets, code, code segments, program code, programs, subprograms, software modules, applications, software applications, software packages, routines, subroutines, objects, executable files, threads of execution, procedures, functions, etc., regardless of whether they are called software, firmware, middleware, microcode, hardware description language, or by other names.

[0054] Software, instructions, information, etc. may also be transmitted or received over a transmission medium. For example, if software is transmitted from a website, server, or other remote source using wired technologies (such as coaxial cable, fiber optic cable, twisted pair, Digital Subscriber Line (DSL)), and / or wireless technologies (such as infrared, microwave), then these wired and / or wireless technologies are included within the definition of transmission media.

[0055] (8) In each of the foregoing embodiments, the terms "system" and "network" are used interchangeably.

[0056] (9) The information, parameters, etc. described in this disclosure may be expressed using absolute values, relative values ​​from a predetermined value, or corresponding other information.

[0057] (10) In the above-described embodiments, the portable device may be a mobile station (MS). Those skilled in the art may also refer to a mobile station as a subscriber station, mobile unit, subscriber unit, wireless unit, remote unit, mobile device, wireless device, wireless communication device, remote device, mobile subscriber station, access terminal, mobile terminal, wireless terminal, remote terminal, handset, user agent, mobile client, client, or some other appropriate term. In this disclosure, terms such as "mobile station," "user terminal," "user equipment (UE)," and "terminal" may be used interchangeably.

[0058] (11) In the above-described embodiments, the terms "connected," "coupled," or any variations thereof refer to any direct or indirect connection or coupling between two or more elements, and may include the presence of one or more intermediate elements between two elements that are "connected" or "coupled" to each other. The coupling or connection between elements may be physical, logical, or a combination thereof. For example, "connected" may be read as "access." As used in this disclosure, two elements may be considered to be "connected" or "coupled" to each other using at least one of one or more wires, cables, and printed electrical connections, as well as electromagnetic energy having wavelengths in the radio frequency range, microwave range, and optical (both visible and invisible) range, as some non-limiting and non-exhaustive examples.

[0059] (12) In the above embodiments, the phrase "based on" does not mean "based only on," unless otherwise specified. In other words, the phrase "based on" means both "based only on" and "based at least on."

[0060] (13) As used in this disclosure, the terms "determining" and "determining" may encompass a wide variety of actions. "Determining" and "determining" may include, for example, judging, calculating, computing, processing, deriving, investigating, looking up, searching, inquiring (e.g., searching a table, database, or other data structure), and ascertaining something that is considered to be a "determining." Also, "determining" and "determining" may include receiving (e.g., receiving information), transmitting (e.g., sending information), input, output, accessing (e.g., accessing data in memory), and the like. Furthermore, "judgment" and "decision" can include regarding resolving, selecting, choosing, establishing, comparing, etc. as having been "judged" or "decided." In other words, "judgment" and "decision" can include regarding some action as having been "judged" or "decided." Furthermore, "judgment (decision)" can be interpreted as "assuming," "expecting," "considering," etc.

[0061] (14) In the above embodiments, when the terms "include," "including," and variations thereof are used, these terms are intended to be inclusive, similar to the term "comprising." Furthermore, the term "or," as used in this disclosure, is not intended to be an exclusive or.

[0062] (15) In this disclosure, where articles are added by translation, such as a, an, and the in English, this disclosure may include that the nouns following these articles are plural.

[0063] (16) In this disclosure, the term "A and B are different" may mean "A and B are different from each other." The term may also mean "A and B are each different from C." Terms such as "separate" and "combined" may also be interpreted in the same way as "different."

[0064] (17) The aspects / embodiments described in this disclosure may be used alone, in combination, or switched depending on the implementation. Notification of predetermined information (e.g., notification that "X is true") is not limited to being explicit, but may be implicit (e.g., not notifying the predetermined information).

[0065] E: Aspects Understood from the Above-described Forms or Modifications Although the present disclosure has been described in detail above, it is clear to those skilled in the art that the present disclosure is not limited to the embodiments described herein. The present disclosure can be implemented in modified and altered forms without departing from the spirit and scope of the present disclosure as defined by the claims. Therefore, the description of the present disclosure is intended to be illustrative and does not have any limiting meaning on the present disclosure. The following aspects can be understood from at least one of the above-described embodiments or modifications.

[0066] An estimation device according to a first aspect of the present disclosure includes a first calculation unit, a second calculation unit, and an estimation unit. The first calculation unit calculates a predicted value of a dependent variable for each of a plurality of areas based on a plurality of explanatory variables including a first explanatory variable. The second calculation unit calculates an index indicating the strength of correlation between the difference between the predicted value and the actual measured value of the dependent variable and each of a plurality of second explanatory variables obtained by subdividing the first explanatory variable. The estimation unit estimates a second explanatory variable that contributes to an increase in the dependent variable by comparing the index calculated by the second calculation unit for each of the plurality of second explanatory variables.

[0067] According to the estimation device of the first aspect, when a first explanatory variable included in a plurality of explanatory variables used to calculate a predicted value of a dependent variable is subdivided into a plurality of second explanatory variables, it is possible to estimate a second explanatory variable among the plurality of second explanatory variables that contributes to an increase in the dependent variable.

[0068] In a second aspect (an example of the first aspect) of the present disclosure, the objective variable may be a variable corresponding to the number of active users of a payment service provided in the plurality of areas, the first explanatory variable may be a variable corresponding to the number of affiliated stores that are stores where the payment service is available, and the plurality of second explanatory variables may be variables corresponding to the number of affiliated stores for each industry. According to the second aspect, it is possible to estimate the industry that contributes to an increase in the number of active users.

[0069] An estimation device according to a third aspect (an example of the first aspect) of the present disclosure may further include an extraction unit that extracts, from the plurality of areas, first areas belonging to a first group in which the actual measured value of the dependent variable is lower than the predicted value calculated by the first calculation unit, and extracts second areas belonging to a second group different from the first group. The second calculation unit in the estimation device of the third aspect calculates the index for the first area and the second area. According to the third aspect, since the index indicating the strength of correlation is calculated for the first area and the second area, the computational processing load required to estimate the second explanatory variable that contributes to an increase in the dependent variable is reduced compared to an aspect in which the index is calculated for all of the plurality of areas.

[0070] 10A, 10B... estimation device, 110... communication device, 120... storage device, 130... processing device, 130a... first calculation unit, 130b, 130e... second calculation unit, 130c... estimation unit, 130d... extraction unit, 140... bus, PR1, PR2... programs.

Claims

1. a first calculation unit that calculates a predicted value of a dependent variable based on a plurality of explanatory variables including a first explanatory variable for each of a plurality of areas; a second calculation unit that calculates an index indicating the strength of correlation between a difference between the predicted value and an actual measured value of the dependent variable and each of a plurality of second explanatory variables obtained by subdividing the first explanatory variable, the second explanatory variables not being used in calculating the predicted value; and an estimation unit that estimates a second explanatory variable that contributes to an increase in the objective variable by comparing the indexes calculated by the second calculation unit for each of the plurality of second explanatory variables; An estimation device having:

2. the objective variable is a variable corresponding to the number of active users of a payment service provided in the plurality of areas, the first explanatory variable is a variable corresponding to the number of affiliated stores that are stores where the payment service can be used, The estimation device according to claim 1 , wherein the plurality of second explanatory variables are variables corresponding to the number of affiliated stores for each business type.

3. an extracting unit that extracts, from the plurality of areas, a first area belonging to a first group in which the actual measured value of the dependent variable is lower than the predicted value calculated by the first calculating unit, and extracts a second area belonging to a second group different from the first group, The second calculation unit calculating the index for the first area and the second area; 2. The estimation device according to claim 1 .