Analytical model creation support device, analytical model creation support method, program, and recording medium

Abstract

To provide an analysis model creation support device, an analysis model creation support method, a program and a recording medium which can efficiently create an analysis model.SOLUTION: An analysis model creation support device includes an analysis object data acquisition unit, a correlation coefficient calculation unit, a group data creation unit, a data pattern creation unit, an analysis model creation unit, a model accuracy calculation unit and an output unit. The analysis object data acquisition unit acquires analysis object data including a plurality of numeric data. The correlation coefficient calculation unit calculates correlation coefficients between numeric data. The group data creation unit creates group data grouping the correlated numeric data, based on the correlation coefficients. The data pattern creation unit generates a data pattern including at least one of the numeric data or the group data. The analysis model creation unit creates an analysis model for each data pattern. The model accuracy calculation unit calculates accuracy of each analysis model. The output unit outputs the accuracy of the analysis model.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an analytical model creation support device, an analytical model creation support method, a program, and a recording medium. [Background technology]

[0002] In various fields, such as electricity demand forecasting, loan feasibility assessment, and plant operation support, analytical models are created based on multiple numerical data to predict target data (Patent Documents 1 and 2, etc.). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2005-332360 [Patent Document 2] Japanese Patent Application Publication No. 2019-144747 Summary of the Invention [Problem to be solved by the invention]

[0004] In creating the analytical model, it is necessary to prepare various data to improve the accuracy of predictions, and then manually select and discard the data to create the analytical model. However, as the amount of data increases, the number of patterns for creating analytical models increases exponentially, and there is a problem in that manually selecting and evaluating the data is time-consuming.

[0005] Therefore, an object of the present invention is to provide an analytical model creation support device that can efficiently create an analytical model. [Means for solving the problem]

[0006] In order to achieve the above object, the analytical model creation support device of the present invention comprises: The system includes an analysis target data acquisition unit, a correlation coefficient calculation unit, a group data creation unit, a data pattern creation unit, an analysis model creation unit, a model accuracy calculation unit, and an output unit, the analysis target data acquisition unit acquires analysis target data including a plurality of numerical data; the correlation coefficient calculation unit calculates a correlation coefficient between the numerical data; the group data creation unit creates group data by grouping correlated numerical data together based on the correlation coefficient; the data pattern creation unit generates a data pattern including at least one of the numerical data and the group data; the analytical model creation unit creates an analytical model for each of the data patterns, the model accuracy calculation unit calculates the accuracy of each analytical model; The output unit outputs the accuracy of the analytical model.

[0007] The analytical model creation support method of the present invention includes: The method includes an analysis target data acquisition step, a correlation coefficient calculation step, a group data creation step, a data pattern creation step, an analysis model creation step, a model accuracy calculation step, and an output step, the analysis target data acquisition step acquires analysis target data including a plurality of numerical data; the correlation coefficient calculation step calculates a correlation coefficient between the numerical data; the group data creation step creates group data by grouping correlated numerical data together based on the correlation coefficient; the data pattern creating step generates a data pattern including at least one of the numerical data and the group data; the analytical model creation step creates an analytical model for each of the data patterns, The model accuracy calculation step calculates the accuracy of each analytical model, The output step outputs the accuracy of the analytical model.

[0008] The program of the present invention is a program for causing a computer to execute an analysis target data acquisition procedure, a correlation coefficient calculation procedure, a group data creation procedure, a data pattern creation procedure, an analytical model creation procedure, a model accuracy calculation procedure, and an output procedure, the step of acquiring data to be analyzed acquires data to be analyzed including a plurality of pieces of numerical data; the correlation coefficient calculation step calculates a correlation coefficient between the numerical data; the group data creation step creates group data by grouping correlated numerical data together based on the correlation coefficient; the data pattern generation step generates a data pattern including at least one of the numerical data and the group data; the analytical model creation step creates an analytical model for each of the data patterns; The model accuracy calculation step calculates the accuracy of each analytical model; The output procedure is a program that outputs the accuracy of the analytical model. [Effects of the Invention]

[0009] According to the present invention, an analytical model can be created efficiently. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 is a block diagram showing an example of the configuration of an analytical model creation support device according to the first embodiment. [Figure 2] FIG. 2 is a block diagram illustrating an example of a hardware configuration of the analytical model creation support device according to the first embodiment. [Figure 3] FIG. 3 is a flowchart illustrating an example of processing in the analytical model creation support device according to the first embodiment. [Figure 4] FIG. 4 is a block diagram illustrating an example of the configuration of an analytical model creation support device according to the second embodiment. [Figure 5] FIG. 5 is a flowchart showing an example of processing in the analytical model creation support device of the second embodiment. [Figure 6] FIG. 6 is a block diagram illustrating an example of the configuration of an analytical model creation support device according to the third embodiment. [Figure 7] FIG. 7 is a flowchart showing an example of processing in the analytical model creation support device according to the third embodiment. [Figure 8] FIG. 8 is a block diagram illustrating an example of the configuration of an analytical model creation support device according to the fourth embodiment. [Figure 9] FIG. 9 is a flowchart showing an example of processing in the analytical model creation support device according to the fourth embodiment. [Figure 10] FIG. 10 is a block diagram illustrating an example of the configuration of an analytical model creation support device according to the fifth embodiment. [Figure 11] FIG. 11 is a flowchart showing an example of processing in the analytical model creation support device according to the fifth embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0011] The analytical model creation support device of the present invention can be used to create an analytical model using, for example, supervised learning, which can include logistic regression, linear regression, decision tree, random forest, elastic net, support vector machine (SVM), regularization, k-nearest neighbor (k-NN), naive Bayes, neural network (NN) perceptron, etc.

[0012] Embodiments of the present invention will be described with reference to the drawings. The present invention is not limited to the following embodiments. In the following drawings, the same parts are denoted by the same reference numerals. Furthermore, the descriptions of the embodiments can be mutually incorporated unless otherwise specified, and the configurations of the embodiments can be combined unless otherwise specified.

[0013] [Embodiment 1] Fig. 1 is a block diagram showing an example of the configuration of an analytical model creation support device 10 according to this embodiment. As shown in Fig. 1, this device 10 includes an analysis target data acquisition unit 11, a correlation coefficient calculation unit 12, a group data creation unit 13, a data pattern creation unit 14, an analytical model creation unit 15, a model accuracy calculation unit 16, and an output unit 17. Although not shown, this device 10 may also include, for example, a storage unit.

[0014] The device 10 may be, for example, a single device including the above-described units, or a device in which the units can be connected via a communication network. The device 10 can also be connected to an external device (described later) via the communication network. The communication network is not particularly limited and any known network can be used, for example, a wired or wireless network. Examples of the communication network include the Internet, the World Wide Web (WWW), a telephone line, a Local Area Network (LAN), a Storage Area Network (SAN), a Delay Tolerant Networking (DTN), a Low Power Wide Area Network (LPWA), and a Local 5G (L5G). Examples of wireless communication include Wi-Fi (registered trademark), Bluetooth (registered trademark), Local 5G, and LPWA. Examples of the wireless communication include direct communication between devices (Ad Hoc communication), infrastructure communication, and indirect communication via an access point. The device 10 may be incorporated into a server as a system. Furthermore, the device 10 may be, for example, a personal computer (PC, for example, a desktop or notebook type) on which the program of the present invention is installed, a smartphone, a tablet terminal, etc. The device 10 may be in the form of cloud computing or edge computing, for example, in which at least one of the above-mentioned units is located on a server and the other units are located on a terminal.

[0015] 2 shows a block diagram of the hardware configuration of the device 10. The device 10 includes, for example, a CPU 101, a memory 102, a bus 103, a storage device 104, an input device 106, a display (display device) 107, and a communication device (communication unit) 108. The components of the device 10 are connected to each other via the bus 103 and their respective interfaces (I / F).

[0016] The CPU 101 cooperates with other components, for example, via a controller (such as a system controller or an I / O controller), and is responsible for overall control of the device 10. In the device 10, the CPU 101 executes, for example, the program 105 of the present invention and other programs, and also reads and writes various types of information. Specifically, for example, the CPU 101 functions as an analysis target data acquisition unit 11, a correlation coefficient calculation unit 12, a group data creation unit 13, a data pattern creation unit 14, an analytical model creation unit 15, a model accuracy calculation unit 16, and an output unit 17. The device 10 includes a CPU as a computing device, but may also include other computing devices such as a GPU (Graphics Processing Unit) or an APU (Accelerated Processing Unit), or may include a combination of the CPU and these.

[0017] The bus 103 can also be connected to, for example, an external device. Examples of the external device include an external storage device (such as an external database), a printer, an external input device, an external display device, and an external imaging device. The device 10 can be connected to an external network (the communication line network) by, for example, a communication device 107 connected to the bus 103, and can also be connected to other devices via the external network.

[0018] The memory 102 may be, for example, a main memory (primary storage device). When the CPU 101 performs processing, the memory 102 reads various operation programs, such as a program 105 of the present invention stored in a storage device 104 (described later), and the CPU 101 receives data from the memory 102 and executes the program. The main memory may be, for example, a RAM (random access memory). The memory 102 may also be, for example, a ROM (read only memory).

[0019] The storage device 104 is also referred to as an auxiliary storage device, for example, in contrast to the main memory (primary storage device). As described above, the storage device 104 stores an operating program including the program of the present invention. The storage device 104 may be, for example, a combination of a recording medium and a drive for reading and writing data from and to the recording medium. The recording medium is not particularly limited and may be, for example, an internal or external type, such as a hard disk (HD), CD-ROM, CD-R, CD-RW, MO, DVD, flash memory, or memory card. The storage device 104 may be, for example, a hard disk drive (HDD) or a solid-state drive (SSD) in which the recording medium and drive are integrated. When the device 10 includes the storage unit, the storage device 104 functions as the storage unit, for example.

[0020] In the present device 10, the memory 102 and the storage device 104 can also store various information such as log information, information acquired from an external database (not shown) or an external device, information generated by the present device 10, and information used when the present device 10 executes processing. Note that at least a portion of the information may be stored, for example, in an external server other than the memory 102 and the storage device 104, or may be stored in a distributed manner across multiple terminals using blockchain technology or the like.

[0021] The device 10 further includes, for example, an input device 106 and a display 107. Examples of the input device 106 include pointing devices such as a touch panel, track pad, and mouse; a keyboard; imaging means such as a camera and scanner; card readers such as an IC card reader and a magnetic card reader; and audio input means such as a microphone. Examples of the display 107 include display devices such as an LED display and a liquid crystal display. In the first embodiment, the input device 106 and the display 107 are configured separately, but the input device 106 and the display 107 may be configured as an integrated unit, such as a touch panel display.

[0022] Next, an example of the analytical model creation support method of this embodiment will be described with reference to the flowchart in Fig. 3. The analytical model creation support method of this embodiment is implemented as follows, for example, using the analytical model creation support device 10 of Fig. 1 or Fig. 2. Note that the analytical model creation support method of this embodiment is not limited to use of the analytical model creation support device 10 of Fig. 1 or Fig. 2. In the following description, as a specific example, a case will be described in which data 1, data 2, data 3, and data 4 exist for each of the target data shown in Table 1 below as analysis target data for analyzing the target data, but the analytical model creation support method of the present invention is not limited or restricted in any way by the following description.

[0023] [Table 1]

[0024] First, the analysis target data acquiring unit 11 acquires analysis target data including a plurality of pieces of numerical data (Data 1 to Data 4) shown in Table 1 (S1, analysis target data acquiring step). The numerical data included in the analysis target data may be a plurality of pieces of numerical data, and there are no particular limitations on the upper limit of the number and type of the numerical data. Specific examples of the analysis target data include temperature, humidity, annual income, and family structure (number of people). The analysis data is not limited to these, and may also be, for example, data obtained by converting text data into variable data. Examples of the text data include weather (sunny, cloudy, rainy, etc.), occupation (national civil servant, company employee, self-employed, etc.), and country (Japan, USA, UK, etc.). The analysis target data acquiring unit 11 may acquire the analysis target data from a database external to the device via the communication network using a communication device 108 connected to the bus 103, or may acquire the analysis target data directly from a sensor or the like external to the device that measures the analysis target data. Alternatively, the analysis target data may be acquired from the memory 102 or storage device 104 of the analytical model creation support device 10 of the present invention.

[0025] Next, the correlation coefficient calculation unit 12 calculates the correlation coefficient between the numerical data (S2, correlation coefficient calculation step). The correlation coefficient between the numerical data can be calculated, for example, according to the following formula (1), by dividing the covariance of two numerical data (variable A and variable B) by the standard deviation of variable A and the standard deviation of variable B. The correlation coefficient r takes a value between -1 and 1, for example. When the correlation coefficient r is a positive value, it can be said that there is a positive correlation between variable A and variable B, and when the correlation coefficient r is a negative value, it can be said that there is a negative correlation between variable A and B. As a specific example, the correlation coefficients of data 1 to data 4 shown in Table 1 are shown in Table 2 below. Correlation coefficient r = covariance / (standard deviation of variable A × standard deviation of variable B)...(1)

[0026] [Table 2]

[0027] Next, the group data creation unit 13 creates group data by grouping correlated numerical data together based on the correlation coefficient (S3, group data creation step). Specifically, the group data creation unit 13 determines that data whose correlation coefficient exceeds a threshold is correlated, and groups the correlated numerical data together. The threshold is not particularly limited and may be a predetermined value or a value set by a user who has confirmed the correlation coefficient. The latter form will be described later in embodiment 2. As a specific example, when the threshold is set to "0.9 or more," the group data creation unit 13 determines that data 1 and data 2 are correlated based on the correlation coefficients shown in Table 2, and creates group data 1 by grouping data 1 and data 2 together as shown in Table 3 below. For example, the group data creation unit 13 may group numerical data whose correlation between the numerical data is positive or negative, but it is preferable to group numerical data whose correlation between the numerical data is positive.

[0028] [Table 3]

[0029] Next, the data pattern creation unit 14 creates a data pattern including at least one of the numeric data and the group data (S4, data pattern creation step). The data pattern creation unit 14 may create data patterns for all combination patterns of the numeric data and the group data, or may create data patterns for some of the combination patterns. As a specific example, if group data 1 is created in S3, the data pattern creation unit 14 creates seven data patterns: combination 1 of group data 1 and data 3; combination 2 of group data 1, data 3, and data 4; combination 3 of group data 1 and data 4; combination 5 of group data 1 only; combination 6 of data 3 only; and combination 7 of data 4 only.

[0030] Next, the analytical model creation unit 15 creates an analytical model for each of the data patterns (S5, analytical model creation step). The analytical model can be created using a method similar to that used by a known analytical model creation AI engine. For example, the analytical model creation unit 15 may use the target data and the data pattern created in S4 to generate the analytical model and analysis results for each of the data patterns, or may acquire the analytical model and analysis results created by a known analytical model creation AI engine external to the device 10. In the latter case, for example, the analytical model creation unit 15 can create the analytical model by outputting the target data and the data pattern created in S4 to an analytical model creation AI engine external to the analytical model creation support device 10 and acquiring the analytical model and analysis results generated by the AI ​​engine for each of the data patterns.

[0031] In S5, when creating an analytical model of a data pattern including the group data, any data can be used as the multiple numerical data included in the group data. For example, the first data of the numerical data included in the group data may be used, or any data specified by the user may be used. The latter form will be described later in embodiment 3. Table 4 below shows an example of an analytical model using the data pattern of combination 2 of group data 1, data 3, and data 4. The model shown in Table 4 below is a prediction result by an analytical model using data 1, which is the first data of the numerical data included in group data 1 in the data pattern of combination 2.

[0032] [Table 4]

[0033] Next, the model accuracy calculation unit 16 calculates the accuracy of each analytical model (S6, model accuracy calculation step). Specifically, the model accuracy calculation unit 16 can calculate the accuracy of the analytical model by, for example, evaluating predicted values ​​(prediction results by the analytical model for each data pattern) against actual measured values ​​(the target data). The method of calculating the accuracy of the analytical model by the model accuracy calculation unit 16 can be determined appropriately depending on, for example, the type of analytical model created and the purpose of the analysis. If the analytical model is, for example, a regression analysis model, methods such as mean error (ME), root mean squared error (RMSE), mean absolute error (MAE), mean percentage error (MPE), mean absolute percentage error (MAPE), mean squared error (MSE), and coefficient of determination (R2) can be used. Furthermore, if the analytical model is, for example, a classification model, methods such as a confusion matrix, accuracy, precision, recall, F-measure, etc. The accuracy calculation method can be appropriately set by, for example, a user of the analytical model creation support device 10 depending on the purpose.

[0034] Table 5 below shows, as a specific example of the model accuracy of the analytical model for each data pattern, the RMSE results calculated by the model accuracy calculation unit 16. The closer the RMSE value is to 0, the smaller the error between the actual measurement value and the predicted value, i.e., the higher the accuracy of the analytical model.

[0035] [Table 5]

[0036] The analytical model creation support device 10 may, for example, repeatedly execute S5 and S6 until creation of analytical models and calculation of accuracy are completed for all data patterns created in S4, or may repeat S5 and S6 until a predetermined condition is met. The predetermined condition may be, for example, an arbitrarily specified number of times, or until the accuracy of the analytical model exceeds a specified accuracy condition. When S5 and S6 are repeated until the predetermined condition is met, it is preferable to select the data patterns evenly so as not to bias the data used in creating the analytical model in S5.

[0037] Then, the output unit 17 outputs the accuracy of the analytical model (S7, output step), and ends (END) the processing of the analytical model creation support device 10. The output unit 17 may, for example, output (display) the accuracy of the analytical model on the display 107, or may output it to an external device via a communication network.

[0038] According to the analytical model creation support device 10 of this embodiment, for example, the correlation coefficient calculation unit 12 calculates the correlation coefficient between numerical data, and the group data creation unit 13 groups correlated numerical data together. Therefore, according to the analytical model creation support device 10, the amount of data used to create an analytical model can be reduced, and analytical models can be created and evaluated efficiently.

[0039] [Embodiment 2] This embodiment is similar to the analytical model creation support device 10 of embodiment 1, and the description therefor can be cited, except that it includes a correlation coefficient output unit and a correlation threshold acquisition unit in addition to the configuration of the analytical model creation support device 10 of embodiment 1. The analytical model creation support device of this embodiment includes a correlation coefficient output unit and a correlation threshold acquisition unit, the correlation coefficient output unit outputs a correlation coefficient between the numerical data, the correlation threshold acquisition unit acquires a correlation threshold between the numerical data, and the group data creation unit determines a correlation between the numerical data based on the correlation coefficient and the correlation threshold, and creates group data by grouping together numerical data determined to be correlated.

[0040] Fig. 4 is a block diagram showing an example of the configuration of an analytical model creation support device 10A of this embodiment. As shown in Fig. 4, the analytical model creation support device 10A includes a correlation coefficient output unit 21 and a correlation threshold acquisition unit 22 in addition to the configuration of the analytical model creation support device 10 of embodiment 1. The hardware configuration of the analytical model creation support device 10A is the same as that of the analytical model creation support device 10 of Fig. 2, except that the CPU 101 includes the configuration of the analytical model creation support device 10A of Fig. 4 instead of the configuration of the analytical model creation support device 10 of Fig. 1.

[0041] Next, the analytical model creation support method of this embodiment will be described with reference to the flowchart of Fig. 5. The analytical model creation support method of this embodiment can be implemented, for example, using an analytical model creation support device 10A of this embodiment shown in Fig. 4. Note that the analytical model creation support method of the present invention is not limited to use of the analytical model creation support device 10A.

[0042] First, steps S1 and S2 are carried out in the same manner as steps S1 and S2 in the analytical model creation support method of the first embodiment.

[0043] Next, the correlation coefficient output unit 21 outputs the correlation coefficient between the numerical data calculated in S2 (S21, correlation coefficient output step). The correlation coefficient output unit 21 may, for example, output (display) the correlation coefficient on the display 107, or may output it to an external device via a communication network.

[0044] Next, the correlation threshold acquisition unit 22 acquires a correlation threshold between the numerical data (S22, correlation threshold acquisition step). The user checks the output correlation coefficient, sets a correlation threshold that will be the threshold for the correlation between the numerical data, and inputs the correlation threshold via the input device 106 of the analytical model creation support device 10A. The correlation threshold acquisition unit 22 acquires, for example, the correlation threshold input by the user.

[0045] The group data creation unit 13 determines the correlation between the numerical data based on the correlation coefficient and the correlation threshold value acquired in S22, and creates group data by grouping the numerical data determined to have the correlation (S3, group data creation step).

[0046] Then, S4 to S7 are carried out in the same manner as S4 to S7 in the analytical model creation support method of the first embodiment, and the processing of the analytical model creation support device 10A is ended (END).

[0047] In the analytical model creation support device of this embodiment, before creating group data, the correlation coefficient between numerical data is output to the user by, for example, the correlation coefficient output unit 21. This allows the user to set a threshold value for determining that there is a correlation between numerical data in accordance with the purpose and method of analysis, thereby enabling more efficient creation of an analytical model.

[0048] [Embodiment 3] This embodiment is similar to the analytical model creation support device 10 of embodiment 1, and the description therefor can be cited, except that it includes a group data output unit and a representative data acquisition unit in addition to the configuration of the analytical model creation support device 10 of embodiment 1. The analytical model creation support device of this embodiment includes a group data output unit and a representative data acquisition unit, the group data output unit outputs the group data, the representative data acquisition unit acquires representative data to be used in creating an analytical model for each of the group data, and the analytical model creation unit creates an analytical model using the representative data as numerical data of the group data in creating an analytical model of a data pattern including the group data.

[0049] Fig. 6 is a block diagram showing an example of the configuration of an analytical model creation support device 10B of this embodiment. As shown in Fig. 6, the analytical model creation support device 10B includes a group data output unit 31 and a representative data acquisition unit 32 in addition to the configuration of the analytical model creation support device 10 of embodiment 1. The hardware configuration of the analytical model creation support device 10B is the same as that of the analytical model creation support device 10 of Fig. 2, except that the CPU 101 includes the configuration of the analytical model creation support device 10B of Fig. 6 instead of the configuration of the analytical model creation support device 10 of Fig. 1.

[0050] Next, the analytical model creation support method of this embodiment will be described with reference to the flowchart of Fig. 7. The analytical model creation support method of this embodiment can be implemented, for example, by using an analytical model creation support device 10B of this embodiment shown in Fig. 6. Note that the analytical model creation support method of the present invention is not limited to use of the analytical model creation support device 10B.

[0051] First, steps S1 to S3 are carried out in the same manner as steps S1 to S3 in the analytical model creation support method of the first embodiment.

[0052] Next, the group data output unit 31 outputs the group data created in S3 (S31, group data output step). The group data output unit 31 may, for example, output (display) the group data on the display 107, or may output the group data to an external device via a communication network.

[0053] Next, the representative data acquisition unit 32 acquires representative data to be used in creating an analytical model for each of the group data (S32, representative data acquisition step). The user checks the output group data, determines representative data to be used in creating an analytical model from the numerical data contained in the group data, and inputs the representative data for each of the group data via the input device 106 of the analytical model creation support device 10B. The representative data acquisition unit 32 acquires the representative data input by the user.

[0054] Next, the data pattern creation unit 14 generates a data pattern including at least one of the numerical data and the group data (S4, data pattern creation step). When creating a data pattern including the group data, the data pattern creation unit 14 may create the data pattern by linking the representative data acquired in S32 with the group data, for example.

[0055] Next, the analytical model creation unit 15 creates an analytical model for each of the data patterns (S5, analytical model creation step). For example, in creating an analytical model for a data pattern including the group data, the analytical model creation unit 15 of the analytical model creation support device 10B of the second embodiment creates an analytical model using the representative data as numerical data of the group data.

[0056] Then, S6 to S7 are carried out in the same manner as S6 to S7 in the analytical model creation support method of the first embodiment, and the processing of the analytical model creation support device 10B is terminated (END).

[0057] The analytical model creation support device of this embodiment outputs group data to the user before creating an analytical model, for example, by the group data output unit 31. This allows the user to select, as representative data, numerical data from the group data to be used to create an analytical model in accordance with the purpose and method of analysis, thereby enabling more efficient creation of an analytical model.

[0058] [Embodiment 4] This embodiment is similar to the analytical model creation support device 10 of embodiment 1, except that it includes a score definition information acquisition unit and a data score calculation unit in addition to the configuration of the analytical model creation support device 10 of embodiment 1, and the description thereof can be cited. The analytical model creation support device of this embodiment includes a score definition information acquisition unit and a data score calculation unit, the score definition information acquisition unit acquires score definition information for the numerical data, the score definition information being information linking the accuracy of the analytical model with the score definition of the data, the data score calculation unit calculates a score for each numerical data used in creating the analytical model based on the accuracy of the analytical model and the score definition information, and the output unit outputs the score of the numerical data.

[0059] Fig. 8 is a block diagram showing an example of the configuration of an analytical model creation support device 10C of this embodiment. As shown in Fig. 8, the analytical model creation support device 10C includes a score definition information acquisition unit 41 and a data score calculation unit 42 in addition to the configuration of the analytical model creation support device 10 of embodiment 1. The hardware configuration of the analytical model creation support device 10C is the same as that of the analytical model creation support device 10 of Fig. 2, except that the CPU 101 includes the configuration of the analytical model creation support device 10C of Fig. 8 instead of the configuration of the analytical model creation support device 10 of Fig. 1.

[0060] Next, the analytical model creation support method of this embodiment will be described with reference to the flowchart of Fig. 9. The analytical model creation support method of this embodiment can be implemented, for example, by using an analytical model creation support device 10C of this embodiment shown in Fig. 8. Note that the analytical model creation support method of the present invention is not limited to use of the analytical model creation support device 10C.

[0061] First, steps S1 to S6 are carried out in the same manner as steps S1 to S6 in the analytical model creation support method of the first embodiment.

[0062] Next, the score definition information acquisition unit 41 acquires score definition information for the numerical data (S41, score definition information acquisition step). The score definition information is, for example, information linking the accuracy of the analytical model with the score definition of the data. Table 6 below shows a specific example of the score definition information. The score definition information shown in Table 6 below is information linking the RMSE calculation result with the data score. Note that the score definition information shown in Table 6 below is an example, and the score definition information can be set appropriately according to, for example, the purpose and method of analysis. In the score definition shown in Table 6, as an example, the lower the RMSE, the higher the score, and the higher the RMSE, the lower the score; in other words, the higher the accuracy of the analytical model, the higher the score of the numerical data, and the lower the accuracy of the analytical model, the lower the score of the numerical data; however, the present invention is not limited to this in any way.

[0063] [Table 6]

[0064] Next, the data score calculation unit 42 calculates a score for each piece of numerical data used to create the analytical model based on the accuracy of the analytical model and the score definition information (S42, data score calculation step). The data score calculation unit 42 calculates the score of the numerical data included in the data pattern used to create each analytical model based on the accuracy of each analytical model and the score definition information, and for each piece of numerical data, calculates the average of the scores of the data patterns including the numerical data as the score of the numerical data. As a specific example, Table 7 below shows the scores of data patterns when the accuracy of the analytical model is the accuracy shown in Table 5 and the score definition information is that shown in Table 6. As shown in Table 7 below, the scores of the data pattern including group data 1 are 0, 6, 2, and 8, so the score of group data 1 can be calculated as 4. Furthermore, the scores of the data pattern including data 3 are 6, 6, 8, and 6, so the score of data 3 can be calculated as 6.5. The scores of the data patterns including data 4 are 4, 2, 8, and 6, so the score of data 4 can be calculated as 5.

[0065] [Table 7]

[0066] Then, the output unit 17 outputs the accuracy of the analytical model and the score of the numerical data (S7, output step), and the processing of the analytical model creation support device 10C ends (END).

[0067] The analytical model creation support device of this embodiment can calculate a score for each piece of numerical data based on the accuracy of the analytical model, for example, by the data score calculation unit 42. Therefore, according to the analytical model creation support device of this embodiment, it is possible to evaluate the validity of the data used to create the analytical model, and to create the analytical model more efficiently.

[0068] [Embodiment 5] This embodiment is similar to the analytical model creation support device 10C of embodiment 4, and the description therefor can be cited, except that it includes a selected data acquisition unit in addition to the configuration of the analytical model creation support device 10C of embodiment 4. The analytical model creation support device of this embodiment includes a selected data acquisition unit, which acquires selected data information, the selected data information including selected data selected by a user based on the score of the numerical data, the data pattern creation unit recreates a data pattern based on the selected data information, the analytical model creation unit creates an analytical model for each of the recreated data patterns, the model accuracy calculation unit calculates the accuracy of each analytical model, and the output unit outputs the accuracy of the analytical model.

[0069] Fig. 10 is a block diagram showing an example of the configuration of an analytical model creation support device 10D of this embodiment. As shown in Fig. 10, the analytical model creation support device 10D includes a selected data acquisition unit 51 in addition to the configuration of the analytical model creation support device 10C of embodiment 4. The hardware configuration of the analytical model creation support device 10D is the same as that of the analytical model creation support device 10 of Fig. 2, except that the CPU 101 includes the configuration of the analytical model creation support device 10D of Fig. 10 instead of the configuration of the analytical model creation support device 10 of Fig. 1.

[0070] Next, the analytical model creation support method of this embodiment will be described with reference to the flowchart of Fig. 11. The analytical model creation support method of this embodiment can be implemented, for example, by using an analytical model creation support device 10D of this embodiment shown in Fig. 10. Note that the analytical model creation support method of the present invention is not limited to use of the analytical model creation support device 10D.

[0071] First, steps S1 to S6, S41 to S42, and S7 are carried out in the same manner as steps S1 to S6, S41 to S42, and S7 in the analytical model creation support method of the fourth embodiment.

[0072] Next, the selected data acquisition unit 51 acquires selected data information (S51, selected data acquisition step). The selected data information includes selected data selected by the user based on the scores of the numerical data. The user checks the scores of the output numerical data, selects data to use as analytical data, and inputs the selected data as selected data information via the input device 106 of the analytical model creation support device 10D. The selected data acquisition unit 51 acquires, for example, the selected data information input by the user. The selected data information may include, for example, essential numerical data that must be used in the analysis, and optional numerical data that can be combined in any way.

[0073] Next, the data pattern creation unit 14 recreates the data pattern based on the selected data information (S52, data pattern recreating step). When the selected data information includes the required numeric data, the data pattern creation unit 14 recreates the data patterns so that, for example, all data patterns include the required numeric data. As a specific example, a case will be described in which the selected data information includes data 3 as required data and data 4 as optional data. In this case, the data pattern creation unit 14 creates two data patterns, for example, combination 1 of only data 3, and combination 2 of data 3 and data 4.

[0074] Next, the analytical model creation unit 15 creates an analytical model for each of the recreated data patterns (S52, analytical model re-creation step).Then, the model accuracy calculation unit 16 calculates the accuracy of each recreated analytical model (S53, model accuracy re-calculation step), and the output unit 17 outputs the accuracy of the analytical model (S54, output step), and the processing of the analytical model creation support device 10D ends (END).

[0075] The analytical model creation support device of this embodiment can recreate data patterns based on, for example, selected data information selected by a user, and output the accuracy of the analytical model for each recreated data pattern. Therefore, the analytical model creation support device of this embodiment can create analytical models more efficiently. When creating an analytical model, the analytical model that is output changes depending not only on the data used but also on the parameters set to create the analytical model. This embodiment can further reduce the number of data patterns used to create the analytical model, making it particularly useful for analysis with changed parameters.

[0076] [Embodiment 6] The program of this embodiment is a program for causing a computer to execute each step of the analytical model creation support method described above. Specifically, the program of this embodiment is a program for causing a computer to execute an analysis target data acquisition procedure, a correlation coefficient calculation procedure, a group data creation procedure, a data pattern creation procedure, an analytical model creation procedure, a model accuracy calculation procedure, and an output procedure.

[0077] the step of acquiring data to be analyzed acquires data to be analyzed including a plurality of pieces of numerical data; the correlation coefficient calculation step calculates a correlation coefficient between the numerical data; the group data creation step creates group data by grouping correlated numerical data together based on the correlation coefficient; the data pattern generation step generates a data pattern including at least one of the numerical data and the group data; the analytical model creation step creates an analytical model for each of the data patterns; The model accuracy calculation step calculates the accuracy of each analytical model; The output step outputs the accuracy of the analytical model.

[0078] Furthermore, the program of this embodiment can also be said to be a program that causes a computer to function as an analysis target data acquisition procedure, a correlation coefficient calculation procedure, a group data creation procedure, a data pattern creation procedure, an analytical model creation procedure, a model accuracy calculation procedure, and an output procedure.

[0079] The program of this embodiment can cite the descriptions of the analytical model creation support device and analytical model creation support method of the present invention. For example, the "procedure" in each of the steps can be read as "processing." The program of this embodiment may be recorded on a computer-readable recording medium. The recording medium is, for example, a non-transitory computer-readable storage medium. The recording medium is not particularly limited, and examples thereof include random access memory (RAM), read-only memory (ROM), hard disk (HD), optical disk, and floppy disk (FD).

[0080] Although the present invention has been described above with reference to the embodiments, the present invention is not limited to the above embodiments. Various modifications that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

[0081] <Additional Notes> Some or all of the above embodiments can be described as, but not limited to, the following supplementary notes. (Appendix 1) The system includes an analysis target data acquisition unit, a correlation coefficient calculation unit, a group data creation unit, a data pattern creation unit, an analysis model creation unit, a model accuracy calculation unit, and an output unit, the analysis target data acquisition unit acquires analysis target data including a plurality of numerical data; the correlation coefficient calculation unit calculates a correlation coefficient between the numerical data; the group data creation unit creates group data by grouping correlated numerical data together based on the correlation coefficient; the data pattern creation unit generates a data pattern including at least one of the numerical data and the group data; the analytical model creation unit creates an analytical model for each of the data patterns, the model accuracy calculation unit calculates the accuracy of each analytical model; The analytical model creation support device, wherein the output unit outputs the accuracy of the analytical model. (Appendix 2) a correlation coefficient output unit and a correlation threshold acquisition unit; the correlation coefficient output unit outputs a correlation coefficient between the numerical data; the correlation threshold acquisition unit acquires a correlation threshold between the numerical data; The analytical model creation support device according to claim 1, wherein the group data creation unit determines a correlation between numerical data based on the correlation coefficient and the correlation threshold, and creates group data by grouping together numerical data that are determined to be correlated. (Appendix 3) A group data output unit and a representative data acquisition unit are included, the group data output unit outputs the group data; the representative data acquisition unit acquires representative data to be used for creating an analytical model for each of the group data; 3. The analytical model creation support device according to claim 1, wherein the analytical model creation unit creates an analytical model using the representative data as numerical data of the group data when creating an analytical model of a data pattern including the group data. (Appendix 4) A score definition information acquisition unit and a data score calculation unit are included, the score definition information acquisition unit acquires score definition information for the numerical data, the score definition information is information linking the accuracy of the analytical model with a score definition of data, the data score calculation unit calculates a score for each piece of numerical data used to create the analytical model based on the accuracy of the analytical model and the score definition information; 4. The analytical model creation support device according to claim 1, wherein the output unit outputs a score of the numerical data. (Appendix 5) A selection data acquisition unit is included, the selected data acquisition unit acquires selected data information; the selected data information includes selected data selected by a user based on the score of the numerical data; the data pattern creation unit recreates a data pattern based on the selected data information; the analytical model creation unit creates an analytical model for each of the recreated data patterns, the model accuracy calculation unit calculates the accuracy of each analytical model; 5. The analytical model creation support device according to claim 4, wherein the output unit outputs the accuracy of the analytical model. (Appendix 6) The method includes an analysis target data acquisition step, a correlation coefficient calculation step, a group data creation step, a data pattern creation step, an analysis model creation step, a model accuracy calculation step, and an output step, the analysis target data acquisition step acquires analysis target data including a plurality of numerical data; the correlation coefficient calculation step calculates a correlation coefficient between the numerical data; the group data creation step creates group data by grouping correlated numerical data together based on the correlation coefficient; the data pattern creating step generates a data pattern including at least one of the numerical data and the group data; the analytical model creation step creates an analytical model for each of the data patterns, The model accuracy calculation step calculates the accuracy of each analytical model, The analytical model creation support method, wherein the output step outputs the accuracy of the analytical model. (Appendix 7) a correlation coefficient output step and a correlation threshold acquisition step; the correlation coefficient output step outputs a correlation coefficient between the numerical data; the correlation threshold value acquisition step acquires a correlation threshold value between the numerical data; The analytical model creation support method according to claim 6, wherein the group data creation step determines the correlation between numerical data based on the correlation coefficient and the correlation threshold, and creates group data by grouping together numerical data that are determined to be correlated. (Appendix 8) A group data output step and a representative data acquisition step are included, The group data output step outputs the group data, the representative data acquisition step acquires representative data to be used for creating an analytical model for each of the group data; The analytical model creation support method according to claim 6 or 7, wherein the analytical model creation step creates an analytical model using the representative data as numerical data of the group data when creating an analytical model of a data pattern including the group data. (Appendix 9) The method includes a score definition information acquisition step and a data score calculation step, the score definition information acquisition step acquires score definition information for the numerical data, the score definition information is information linking the accuracy of the analytical model with a score definition of data, the data score calculation step calculates a score for each piece of numerical data used to create the analytical model based on the accuracy of the analytical model and the score definition information; 9. The analytical model creation support method according to any one of appendices 6 to 8, wherein the output step outputs a score of the numerical data. (Appendix 10) A selection data acquisition step is included, The selected data acquisition step acquires selected data information, the selected data information includes selected data selected by a user based on the score of the numerical data; the data pattern creating step recreates a data pattern based on the selected data information; the analytical model creation step creates an analytical model for each of the recreated data patterns, The model accuracy calculation step calculates the accuracy of each analytical model, 10. The analytical model creation support method according to claim 9, wherein the output step outputs the accuracy of the analytical model. (Appendix 11) A program for causing a computer to execute an analysis target data acquisition procedure, a correlation coefficient calculation procedure, a group data creation procedure, a data pattern creation procedure, an analytical model creation procedure, a model accuracy calculation procedure, and an output procedure, the step of acquiring data to be analyzed acquires data to be analyzed including a plurality of pieces of numerical data; the correlation coefficient calculation step calculates a correlation coefficient between the numerical data; the group data creation step creates group data by grouping correlated numerical data together based on the correlation coefficient; the data pattern generation step generates a data pattern including at least one of the numerical data and the group data; the analytical model creation step creates an analytical model for each of the data patterns; The model accuracy calculation step calculates the accuracy of each analytical model; The output step outputs the accuracy of the analytical model. (Appendix 12) A correlation coefficient output procedure and a correlation threshold value acquisition procedure are included. the correlation coefficient output step outputs a correlation coefficient between the numerical data; the correlation threshold value acquisition step acquires a correlation threshold value between the numerical data; The program described in Appendix 11, wherein the group data creation procedure determines the correlation between numerical data based on the correlation coefficient and the correlation threshold, and creates group data by grouping numerical data determined to be correlated. (Appendix 13) It includes a group data output procedure and a representative data acquisition procedure, the group data output step outputs the group data; the representative data acquisition step acquires representative data to be used for creating an analytical model for each of the group data; 13. The program according to claim 11 or 12, wherein the analytical model creation procedure creates an analytical model of a data pattern including the group data by using the representative data as numerical data of the group data. (Appendix 14) Including a procedure for obtaining score definition information and a procedure for calculating data scores, the score definition information acquisition step acquires score definition information for the numerical data, the score definition information is information linking the accuracy of the analytical model with a score definition of data, the data score calculation step calculates a score for each piece of numerical data used to create the analytical model based on the accuracy of the analytical model and the score definition information; 14. The program according to any one of appendices 11 to 13, wherein the output step outputs a score of the numerical data. (Appendix 15) including a selective data acquisition procedure; The selection data acquisition step acquires selection data information, the selected data information includes selected data selected by a user based on the score of the numerical data; the data pattern generation step regenerates a data pattern based on the selected data information; the analytical model creation step creates an analytical model for each of the recreated data patterns; The model accuracy calculation step calculates the accuracy of each analytical model; 15. The program of claim 14, wherein the output step outputs the accuracy of the analytical model. (Appendix 16) A computer-readable recording medium having recorded thereon a program according to any one of appendices 11 to 15. [Industrial Applicability]

[0082] According to the present invention, an analytical model can be efficiently created by reducing the amount of numerical data used to create the analytical model. Therefore, the present invention is useful in various fields where analysis is performed using numerical data, for example. [Explanation of symbols]

[0083] 10, 10A, 10B, 10C, 10D Analysis model creation support device 11 Analysis target data acquisition section 12 Correlation coefficient calculation section 13 Group Data Creation Department 14 Data pattern creation section 15 Analysis Model Creation Department 16 Model accuracy calculation section 17 Output section 21 Correlation coefficient output section 22 Correlation threshold acquisition unit 31 Group data output section 32 Representative data acquisition section 41 Score definition information acquisition unit 42 Data score calculation section 51 Selected data acquisition unit 101 CPU 102 memory 103 Bus 104 Storage device 105 Programs 106 Input Device 107 Display device 108 Communication Devices

Claims

1. The system includes an analysis target data acquisition unit, a correlation coefficient calculation unit, a group data creation unit, a data pattern creation unit, an analysis model creation unit, a model accuracy calculation unit, and an output unit, the analysis target data acquisition unit acquires analysis target data including a plurality of numerical data; the correlation coefficient calculation unit includes a correlation coefficient output unit and a correlation threshold acquisition unit, the correlation coefficient output unit outputs a correlation coefficient between the numerical data; the correlation threshold acquisition unit acquires a correlation threshold between the numerical data; the group data creation unit determines a correlation between the numerical data based on the correlation coefficient and the correlation threshold, and creates group data by grouping together the numerical data determined to be correlated; the data pattern creation unit creates a data pattern of the group data, the analytical model creation unit creates an analytical model for each of the data patterns, the model accuracy calculation unit calculates the accuracy of each analytical model; The analytical model creation support device, wherein the output unit outputs the accuracy of the analytical model.

2. A group data output unit and a representative data acquisition unit are included, the group data output unit outputs the group data; the representative data acquisition unit acquires representative data to be used for creating an analytical model for each of the group data; 2. The analytical model creation support device according to claim 1, wherein the analytical model creation unit creates the analytical model using the representative data as the numerical data of the group data in creating the analytical model of the data pattern including the group data.

3. A score definition information acquisition unit and a data score calculation unit are included, the score definition information acquisition unit acquires score definition information for the numerical data, the score definition information is information linking the accuracy of the analytical model with a score definition of data, the data score calculation unit calculates a score for each piece of numerical data used to create the analytical model based on the accuracy of the analytical model and the score definition information; The analytical model creation support device according to claim 1 , wherein the output unit outputs a score of the numerical data.

4. A selection data acquisition unit is included, the selected data acquisition unit acquires selected data information; the selected data information includes selected data selected by a user based on the score of the numerical data; the data pattern creation unit recreates a data pattern based on the selected data information; the analytical model creation unit creates an analytical model for each of the recreated data patterns, the model accuracy calculation unit calculates the accuracy of each analytical model; The analytical model creation support device according to claim 3 , wherein the output unit outputs the accuracy of the analytical model.

5. A human resource development support method in which each step including an analysis target data acquisition step, a correlation coefficient calculation step, a group data creation step, a data pattern creation step, an analysis model creation step, a model accuracy calculation step, and an output step is executed by a computer, the analysis target data acquisition step acquires analysis target data including a plurality of numerical data; the correlation coefficient calculation step calculates a correlation coefficient between the numerical data; the group data creation step creates group data by grouping correlated numerical data together based on the correlation coefficient; the data pattern creating step generates a data pattern including the group data; the analytical model creation step creates an analytical model for each of the data patterns, The model accuracy calculation step calculates the accuracy of each analytical model, The analytical model creation support method, wherein the output step outputs the accuracy of the analytical model.

6. A program for causing a computer to execute an analysis target data acquisition procedure, a correlation coefficient calculation procedure, a group data creation procedure, a data pattern creation procedure, an analytical model creation procedure, a model accuracy calculation procedure, and an output procedure, the step of acquiring data to be analyzed acquires data to be analyzed including a plurality of pieces of numerical data; the correlation coefficient calculation step calculates a correlation coefficient between the numerical data; the group data creation step creates group data by grouping correlated numerical data together based on the correlation coefficient; the data pattern creation step generates a data pattern including the group data; the analytical model creation step creates an analytical model for each of the data patterns; The model accuracy calculation step calculates the accuracy of each analytical model; The output step outputs the accuracy of the analytical model.

7. A computer-readable recording medium on which the program according to claim 6 is recorded.

Images