Apparatus, program, and method for predicting physical properties of composite materials

Abstract

To perform high-accurate prediction of a physical property value corresponding to a required physical property.SOLUTION: A physical property value prediction device 1 includes: a data creation part 21 for dividing a data set 32 a plurality of times, and creating a plurality of data groups for learning and a plurality of data groups for evaluation; a first regression model creation part 22 for creating a plurality of first regression models on the basis of the plurality of data groups for learning for each of regression methods; a regression method determination part 23 for evaluating predication accuracy by the plurality of first regression models for each of the regression methods on the basis of the plurality of data groups for evaluation, and determining the regression method having the highest evaluation among the plurality of regression methods; a second regression model creation part 24 for creating a second regression model on the basis of the data set, using the regression method having the highest evaluation; and a physical property value prediction part 25 for predicting a physical property value on the basis of the input material data, using the second regression model.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a device, a program, and a method for predicting physical property values ​​of a composite material. [Background technology]

[0002] In recent years, a material design system has been proposed for designing composite materials using the learning results of machine learning (see, for example, Patent Document 1).

[0003] The materials design system described in Patent Document 1 comprises a data file containing a dataset of design conditions and material property values ​​of the material to be designed, and a model learning unit that learns a model for estimating material property values ​​from the design conditions based on the dataset to create a trained model, and verifies the predictive accuracy of the trained model by cross-validation. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2020-77257 Summary of the Invention [Problem to be solved by the invention]

[0005] In recent years, composite materials have been developed that combine multiple types of resins and compounding agents to impart new physical properties to the physical properties of the resin itself. The development of new composite materials requires adjusting the composition ratio of each component material until the composite material has the desired physical properties. Therefore, the development of composite materials requires enormous costs and time. Therefore, from the perspective of improving the efficiency of composite material development, it is desirable to be able to predict, to some extent, the physical properties of the composite material to be tested at the experimental planning stage.

[0006] Furthermore, in the case of composite materials such as wire coating materials, many types of compounding agents are added to the resin, and the physical properties can vary significantly depending on the composition ratio of the materials. This makes it difficult to predict the physical properties of composite materials, and there is a need for technology that can predict these physical properties with high accuracy.

[0007] When creating and evaluating a regression model using a technology that predicts physical properties using a regression model generated by machine learning on a dataset consisting of multiple data sets including materials and physical properties, a hold-out method is generally used if the dataset contains a large number of data sets (e.g., 100,000 or more), in which the process of generating training data and test data is carried out only once to create and evaluate the regression model. On the other hand, if the dataset contains a small number of data sets (e.g., less than 100), a leave-one-out method is used, in which one data set is used as test data and the remaining data as training data, and this process is carried out for each data set to create and evaluate a regression model. However, these methods are not appropriate when the number of data sets is somewhere between the two, or when the dataset is sparse.

[0008] An object of the present invention is to provide a property value prediction device, a property value prediction program, and a property value prediction method for a composite material that can make highly accurate predictions of property values ​​corresponding to required properties. [Means for solving the problem]

[0009] In order to solve the above-mentioned problems, the present invention provides a data generation unit that divides a data set, which includes material data indicating blending information of materials constituting a composite material and physical property data indicating physical property values ​​of physical properties related to the composite material, for each of a plurality of types of composite materials a plurality of times so that the combinations of the data sets are different from each other, thereby generating a plurality of training data sets and a plurality of evaluation data sets; a first regression model creation unit that creates a plurality of first regression models for each of a plurality of predetermined regression methods based on the plurality of learning data groups; a regression method determination unit that evaluates the prediction accuracy of the plurality of first regression models for each of the regression methods based on the plurality of evaluation data groups, and determines the regression method with the highest evaluation among the plurality of regression methods; a second regression model creation unit that creates a second regression model based on a plurality of the data sets included in the data set using the regression method with the highest evaluation; and a physical property value prediction unit that predicts physical property values ​​based on input material data using the second regression model. [Effects of the Invention]

[0010] According to the present invention, it is possible to predict with high accuracy the physical property values ​​corresponding to the required physical properties. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a block diagram showing a schematic configuration example of a physical property prediction device according to an embodiment of the present invention; [Figure 2] FIG. 10 is a diagram illustrating an example of a data set. [Figure 3] FIG. 1A is a conceptual diagram for explaining an example of data generation processing by a data generation unit, and FIG. 1B is a conceptual diagram for explaining data generation processing by a leave-one-out method as a comparative example. [Figure 4] FIG. 10 is a diagram for explaining an example of a first regression model creation process performed by a first regression model creation unit. [Figure 5] 10 is a diagram illustrating an example of a prediction accuracy calculation process performed by an error calculation unit. FIG. [Figure 6] FIG. 10 is a diagram for explaining an example of a second regression model creation process performed by a second regression model creation unit. [Figure 7] FIG. 2 is a block diagram for explaining an example of the overall flow of a physical property value prediction device. [Figure 8] 10 is a flowchart illustrating an example of the overall operation of the physical property value prediction device. DETAILED DESCRIPTION OF THE INVENTION

[0012] [Embodiment Mode] 1 is a block diagram showing a schematic configuration example of a physical property prediction device according to an embodiment of the present invention. This physical property prediction device 1 is composed of a computer or the like, and includes a prediction processing unit 2, a storage unit 3, an input unit 4, and a display unit 5, and has a physical property prediction function of predicting physical property values ​​of composite materials using a regression model.

[0013] The composite materials targeted by the physical property value prediction device 1 are manufactured by combining materials that constitute the composite material. Examples of composite materials include polymer compositions manufactured by combining multiple resins and multiple compounding agents. Examples of polymer compositions include wire coating materials, sheets, tubes, bonded magnets, and magnet rolls.

[0014] In this specification, physical properties refer to properties required of a composite material, and may be, for example, mechanical properties such as tensile breaking elongation and tensile strength, or thermal properties such as heat resistance temperature and thermal expansion coefficient.

[0015] (Explanation of the physical property prediction function) The physical property prediction function of the physical property prediction device 1 selects from a plurality of predetermined regression methods the regression method with the highest prediction accuracy suitable for the predetermined physical properties of the composite material, since the prediction accuracy of the regression method of the regression model used for prediction varies depending on the physical properties, creates a regression model corresponding to the selected regression method by learning it using a dataset, and predicts the physical property values ​​of a new composite material using the learned regression model.

[0016] Furthermore, in order to be able to predict physical property values ​​with high accuracy when the number of data in a dataset is insufficient, the dataset is divided into a training data group and an evaluation data group multiple times to create multiple regression models for each regression method, the average prediction accuracy of each regression model is set as the prediction accuracy of the regression method, and the regression method with the highest prediction accuracy is selected. The division of datasets will be described later.

[0017] The prediction processing unit 2 is composed of a CPU (Central Processing Unit), ROM (Read Only Memory), RAM (Random Access Memory), an interface, etc. The prediction processing unit 2 reads a physical property value prediction program 31 stored in the storage unit 3 into the RAM and executes it with the CPU, thereby functioning as a data generation unit 21, a first regression model creation unit 22, a regression method determination unit 23, a second regression model creation unit 24, and a physical property value prediction unit 25. The regression method determination unit 23 has an error calculation unit 231 and a regression method output unit 232. Details of the units 21 to 25 of the prediction processing unit 2 will be described later. Note that the units 21 to 25 of the prediction processing unit 2 may be composed of hardware such as an FPGA (Field Programmable Gate Array) or an ASIC (Application Specific Integrated Circuit).

[0018] The storage unit 3 is composed of a ROM, a RAM, a hard disk, etc., and stores various types of information such as a physical property prediction program 31 for executing the physical property prediction method according to this embodiment, a dataset 32 ​​(see FIG. 2), a group of regression techniques 33, a training dataset 34, and an evaluation dataset 35.

[0019] The input unit 4 is configured, for example, by a keyboard, a mouse, etc. The display unit 5 is configured, for example, by a liquid crystal display, etc. Note that a touch display in which the input unit 4 and the display unit 5 are integrated may also be used. Furthermore, a terminal device having the input unit 4 and the display unit 5 may be connected to the prediction processing unit 2 via a network.

[0020] Dataset 32 ​​includes data for multiple polymer compositions that are used to create and evaluate regression models. Dataset 32 ​​may contain sparse data, for example, and the number of data items may be insufficient. Details of Dataset 32 ​​will be described later.

[0021] The regression method group 33 includes a plurality of (for example, m) predetermined regression methods RM1 to RM m These are each linked to identification information of the regression method (for example, the name of the regression method, etc.) and stored in the storage unit 3. Examples of regression methods included in the regression method group 33 include a Gaussian process regression method, a linear regression method, a hierarchical Bayes method, and a regression method using a neural network model. A regression method is an algorithm that uses machine learning to learn parameters of a regression model so that a target variable can be predicted from an explanatory variable.

[0022] The training data set 34 is a set of multiple (for example, n) training data groups LDG1 to LDG2, each of which is made up of multiple training data LD generated by the data generation unit 21. n (See Figure 3).

[0023] The evaluation data set 35 includes a plurality of (for example, n) evaluation data groups TDG1 to TDG2, each of which is made up of a plurality of evaluation data TD generated by the data generating unit 21. n (See Figure 3).

[0024] (Dataset configuration) FIG. 2 is a diagram showing an example of dataset 32. Dataset 32 ​​has a data set 321 for each of multiple types of polymer compositions. Specifically, dataset 32 ​​includes a composition ID 320 that identifies the polymer composition, and a data set 321 that is provided corresponding to composition ID 320 and includes material data 321a that indicates blending information of materials that make up the polymer composition and physical property data 321b that indicates physical property values ​​of the polymer composition. FIG. 2 shows an example of nine data sets 321, but the number of data sets 321 (= the number of data) is not limited to nine.

[0025] FIG. 2 illustrates an example of a polymer composition, such as a wire coating material. The wire coating material includes multiple resins, such as “Polymer 1,” “Polymer 2,” “Polymer 3,” “Polymer 4,” and “Polymer 5,” and multiple compounding ingredients, such as “Filler 1,” “Filler 2,” “Filler 3,” and “Crosslinking Agent.” The physical property data 321b indicates, for example, a physical property value of “elongation” (e.g., tensile breaking elongation) as an example of a predetermined physical property among multiple physical properties possessed by the wire coating material. When multiple physical property values ​​are to be predicted, the physical property data 321b may contain multiple physical property values, and a data set 32 ​​may be prepared for each physical property value. For example, when predicting tensile breaking elongation and tensile strength, a regression method suitable for tensile breaking elongation and a regression method suitable for tensile strength are selected from the regression method group 33.

[0026] The blending ratio of "Polymer 1" to "Polymer 5" is, for example, 15 to 45 parts by mass per 100 parts by mass of polymer. The blending ratio of "Filler 1" to "Filler 3" is, for example, 120 to 160 parts by mass per 100 parts by mass of polymer. The blending ratio of "crosslinking agent" is, for example, 0 to 90 parts by mass per 100 parts by mass of polymer. Note that blending ingredients such as fillers and crosslinking agents may be expressed in terms of blending amounts. Blending ratios and blending amounts are examples of blending information.

[0027] Examples of resins include polyolefins such as high-density polyethylene, low-density polyethylene, and ethylene-acrylic acid copolymers, as well as elastomers such as chlorinated polyethylene. Examples of compounding agents include the crosslinking agents shown in Figure 2, as well as fillers such as talc, calcium carbonate, and silica, plasticizers, stabilizers, and the like. It is assumed that the polymer composition will contain 20 or more types of resins, compounding agents, and other materials, but this is not limited to this number.

[0028] The blank rate of the data excluding the composition ID in dataset 32 ​​is assumed to be sparse, at 30% or more and 80% or less, to facilitate the preparation of dataset 32. Here, the "blank rate" refers to the ratio of the number of cells in dataset 32 ​​into which no data has been entered to the total number of cells into which data can be entered. For example, in the case shown in FIG. 2, when the compounding agent is a filler only, the total number of cells is 90, and the number of cells into which no data has been entered is 36, resulting in a blank rate of 40%. Furthermore, the number of data sets 321 (= the number of data) in dataset 32 ​​is assumed to be 100 or more and less than 100,000 (e.g., 350 pieces of data), but is not limited to this.

[0029] Next, the units 21 to 25 that make up the prediction processing unit 2 will be described in detail.

[0030] (Configuration of data generation unit) FIG. 3(a) is a conceptual diagram for explaining an example of data generation processing by the data generator 21, and FIG. 3(b) is a conceptual diagram for explaining data generation processing by the leave-one-out method as a comparative example.

[0031] The data generation unit 21 divides the data set 32 ​​multiple times (n times) so that the combinations of the data sets 321 are different from one another, and generates n training data groups LDG1 to LDG2, each consisting of multiple pieces of training data LD (seven in FIG. 3). n and n evaluation data groups TDG1 to TDG2 each consisting of multiple evaluation data TDs (two in FIG. 3). n The division method may be random or may be a method according to a certain rule, as long as the combinations of data sets 321 are different from each other. "The combinations of data sets 321 are different from each other" means that there is no training data group in which all of the multiple evaluation data sets after division are the same.

[0032] The data generation unit 21 generates n training data groups LDG1 to LDG n are stored in the storage unit 3 as a learning data set 34, and the n evaluation data sets TDG1 to TDGn is stored in the storage unit 3 as an evaluation data set 35.

[0033] Furthermore, the ratio of the number of training data included in the training data group generated by the data generation unit 21 to the number of evaluation data included in the evaluation data group is preferably in the range of, for example, 7:3 to 8:2, taking into consideration the balance between the prediction accuracy and evaluation accuracy of the regression model. The number of divisions is preferably set to be greater as the blank rate of the data increases, in order to suppress bias in the data after division. For example, if the blank rate of the data is 30% or more, the number of divisions is preferably 30 or more.

[0034] If the number of data points in dataset 32 ​​is small (e.g., less than 100,000), or if dataset 32 ​​is sparse (e.g., with a blank space rate of 30% or more), creating a single regression model using all of the data without dividing dataset 32 ​​may result in overfitting of the regression model to dataset 32, potentially resulting in lower prediction accuracy for data other than dataset 32. Furthermore, the hold-out method, in which a regression model is created and evaluated by dividing a dataset only once, is prone to data bias, potentially resulting in lower prediction accuracy or evaluation accuracy for the regression model. Furthermore, the leave-one-out method, as shown in FIG. 3(b), divides dataset 32 ​​into multiple training data groups and one evaluation data group. This increases the number of training data groups, thereby improving the accuracy of the regression model. However, since there is only one evaluation data group, this method may result in lower evaluation accuracy for the regression model. Furthermore, the increased number of divisions results in longer processing time.

[0035] Therefore, in this embodiment, as shown in FIG. 3(a), in order to obtain high prediction accuracy and evaluation accuracy of the regression model without lengthening the processing time, a cross-validation method is adopted in which the dataset 32 ​​is divided multiple times to create multiple regression models (hereinafter also referred to as "first regression models"), and the errors of each first regression model are averaged to evaluate the prediction accuracy of the regression method. Furthermore, since the prediction accuracy of regression methods varies depending on physical properties, this embodiment evaluates the prediction accuracy of each regression method using the cross-validation method for each of the multiple regression methods, selects the regression method with the highest evaluation, and creates a regression model (hereinafter also referred to as "second regression model") corresponding to the regression method with the highest evaluation using all of the datasets. Note that, for example, a random forest may be used as a method for dividing the dataset 32 ​​into evaluation data and training data.

[0036] (Configuration of the first regression model creation unit) The first regression model creation unit 22 uses a plurality of regression methods RM1 to RM m and using the acquired regression method, n training data groups LDG1 to LDG2 acquired from the training data set 34 are acquired. n Based on n first regression models 1M1 to 1M n That is, the first regression model creation unit 22 creates one first regression model corresponding to the regression method by learning a regression model using material data of a plurality of learning data included in one learning data group as explanatory variables and physical property values ​​as objective variables. This is also performed for other learning data groups to create n first regression models 1M1 to 1M for one regression method. n This is then repeated for other regression methods to finally create m × n first regression models 1M1 to 1M n Create a.

[0037] (Configuration of regression method determination unit) The regression method determination unit 23 has an error calculation unit 231 and a regression method output unit 232, and outputs the N first regression models 1M1 to 1M2 created by the first regression model creation unit 22. nThe prediction accuracy of the N evaluation data sets TDG1 to TDG n Based on this, we evaluate each regression method and compare multiple regression methods RM1 to RM m The regression method determination unit 23 determines the regression method RMx with the highest evaluation among them. The regression method determination unit 23 will be specifically described below.

[0038] The error calculation unit 231 calculates, for each regression method and each first regression model, a first average error rate 1MPE1 to 1MPE2 obtained by averaging the errors of the first regression model based on the plurality of evaluation data included in the evaluation data group. n By averaging the errors due to the first regression model, the influence of data bias among the evaluation data can be suppressed.

[0039] Here, the first average error rate 1MPEi (i = 1 to n) is the average value of the error rates PEi (i = 1 to k) calculated for each evaluation data TD for multiple (e.g., k) evaluation data TD included in the evaluation data group TDG, as shown in the following equation (1). 1MPEi=(PE1+PE2+···+PE k ) / k ···(1)

[0040] The error rate PEi (i = 1 to k) is the difference (= error) expressed in percentage between the predicted value PV by the first regression model 1M and the measured value MV, which is a physical property value included in the evaluation data TD, as shown in the following formula (2): The predicted value PV is a value calculated using the first regression model 1M with the material data of the evaluation data TD as an explanatory variable. PEi=(|PV-MV| / MV)×100 ···(2)

[0041] Furthermore, the error calculation unit 231 calculates the first regression models 1M1 to 1M2 as shown in the following equation (3): n The first average error rate of each of 1MPE1 to 1MPE nThe second mean error rate 2MPEi (i = 1 to m) obtained by averaging these is calculated as the prediction accuracy of the regression method RMi (i = 1 to m). Here, the first mean error rate is an example of a first prediction error. The second mean error rate is an example of a second prediction error. Note that the prediction error is not limited to the mean error rate, and may be a mean square error, a mean square deviation, or the like. 2MPEi=(1MPE1+1MPE2+...+1MPE n ) / n ···(3) By calculating the second average error rate by averaging the first average error rates, the influence of data bias among the training data groups can be suppressed.

[0042] (Configuration of the regression method output section) The regression method output unit 232 outputs the regression method RM1 to RM2. m The second average error rates 2MPE1 to 2MPE were calculated as the prediction accuracy for m and determine the regression method RMx with the highest prediction accuracy that corresponds to the smallest second average error rate. The regression method output unit 232 outputs identification information 232a of the determined regression method RMx to the second regression model creation unit 24. This makes it possible to determine a regression method that is suitable for predetermined physical properties.

[0043] (Configuration of the second regression model creation unit) The second regression model creation unit 24 creates a second regression model 2M based on all data sets 321 included in the dataset 32 ​​using the regression method RMx determined by the regression method determination unit 23. That is, the second regression model creation unit 24 uses all data sets 321 included in the dataset 32 ​​as a training data group, and uses the material data of the training data included in the training data group as explanatory variables and the physical property values ​​as target variables to create a second regression model 2M corresponding to the regression method RMx. The second regression model creation unit 24 outputs the calculated second regression model 2M to the physical property value prediction unit 25. By using all data sets 321 included in the dataset 32 ​​for machine learning, it is possible to create a second regression model that can predict physical property values ​​with high accuracy even if the dataset does not have a sufficient number of data and the data is sparse. Note that not all data in the dataset 32 ​​does not need to be used for machine learning; instead, a number of pieces of training data (e.g., 90% or more of the total) greater than the number of pieces of training data when the dataset 32 ​​is divided (e.g., 70% or 80% of the total) may be used for machine learning.

[0044] (Configuration of the physical property prediction unit) The physical property value prediction unit 25 uses the second regression model 2M created by the second regression model creation unit 24 to predict the physical property values ​​of the new polymer composition based on the material data indicating the blending information of the materials that make up the new polymer composition input from the input unit 4, and displays the physical property values ​​on the display unit 5. The physical property values ​​may be displayed on the display unit 5 together with the input material data.

[0045] <Operation of the physical property prediction device> Next, an example of the operation of the physical property value prediction device 1 will be described with reference to Fig. 3 to Fig. 8. Fig. 7 is a block diagram for explaining an example of the overall flow of the physical property value prediction device 1. Fig. 8 is a flowchart for explaining an example of the overall operation of the physical property value prediction device 1.

[0046] (1) Data generation process (S1) This data generation process (S1) will be described with reference to the above-mentioned Fig. 3(a), which is a conceptual diagram for explaining an example of the data generation process by the data generation unit 21.

[0047] In this data generation process (S1), as shown in FIG. 3(a), the data generation unit 21 divides the data set 32 ​​n times so that the combinations of the data sets 321 are different from one another, and generates a plurality of training data groups LDG1 to LDG2, each of which is made up of a plurality of training data (seven in FIG. 3). n and multiple evaluation data groups TDG1 to TDG2 each consisting of multiple evaluation data (two in Figure 3). n and generate.

[0048] Next, the data generation unit 21 generates n training data groups LDG1 to LDG n are stored in the storage unit 3 as a learning data set 34, and n evaluation data sets TDG1 to TDG n is stored in the storage unit 3 as an evaluation data set 35.

[0049] (2) First regression model creation process (S2) This first regression model creation process (S2) will be described with reference to Fig. 4 and Fig. 7. Fig. 4 shows a process in which the first regression model creation unit 22 creates n first regression models 1M1 to 1M n This shows the process of creating the

[0050] In this first regression model creation process (S2), the first regression model creation unit 22 acquires one regression method RM1 from the regression method group 33, acquires a training data group LDG1 from the training dataset 34, and creates a first regression model 1M1 corresponding to the regression method RM1 based on the training data group LDG1, as shown in Figures 4 and 7. Next, the first regression model creation unit 22 acquires a training data group LDG2 from the training dataset 34, and creates a first regression model 1M2 corresponding to the regression method RM1 based on the training data group LDG2. Other training data groups LDG3 to LDG nRegarding this, the first regression model creation unit 22 also uses the other learning data groups LDG3 to LDG n Based on this, the first regression model 1M3 to 1M corresponding to the regression method RM1 n Create a.

[0051] The first regression model creation unit 22 performs the above process using other regression methods RM2 to RM m , and a total of m × n first regression models 1M1 to 1M n The first regression model creating unit 22 creates the m×n first regression models 1M1 to 1M n are linked to the identification information of the corresponding regression model and output to the regression method determination unit 23.

[0052] (3) Prediction accuracy calculation process (S3) This prediction accuracy calculation process (S3) will be described with reference to Fig. 5(a), Fig. 5(b) and Fig. 7. Fig. 5(a) shows a process in which the error calculation unit 231 calculates the first regression models 1M1 to 1M n 5(b) shows the process by which the error calculation unit 231 calculates the prediction accuracy of the second average error rates 2MPE1 to 2MPE m This shows the process of calculating

[0053] As shown in FIG. 5(a), the error calculation unit 231 calculates n first regression models 1M1 to 1M2 created in accordance with one regression method. n The first average error rate is 1MPE1 to 1MPE as the prediction accuracy. n That is, the error calculation unit 231 calculates a first average error rate 1MPE1 of the first regression model 1M1 corresponding to the regression method RM1 based on the evaluation data group TDG1. Next, the error calculation unit 231 calculates a first average error rate 1MPE2 of the first regression model 1M2 corresponding to the regression method RM1 based on the evaluation data group TDG2. n Regarding this, the error calculation unit 231 calculates the other evaluation data groups TDG3 to TDG in the same manner as described above. n Based on this, the first regression model 1M3 to 1M corresponding to the regression method RM1n The first average error rate of 1MPE3 to 1MPE n Calculate.

[0054] The error calculation unit 231 performs the above process using other regression methods RM2 to RM m The first average error rate 1MPE1 to 1MPE is calculated for a total of m × n. n Calculate.

[0055] Next, as shown in FIG. 5(b), the error calculation unit 231 calculates the first average error rates 1MPE1 to 1MPE2 calculated for one regression method RM1. n The second mean error rate 2MPE1 is calculated by averaging the other regression methods RM2 to RM m Similarly, the error calculation unit 231 calculates the second average error rates 2MPE2 to 2MPE m The error calculation unit 231 calculates the m calculated second average error rates 2MPE1 to 2MPE2 as shown in FIG. 5(b) and FIG. m is linked to the identification information of the corresponding regression method and output to the regression method output unit 232.

[0056] (4) Deciding on the regression method (S4) In determining the regression method (S4), the regression method output unit 232 calculates the m number of second average error rates 2MPE1 to 2MPE2 output from the error calculation unit 231 as shown in FIG. m and determines the regression method RMx with the highest prediction accuracy that corresponds to the smallest second average error rate. The regression method output unit 232 outputs identification information 232a of the determined regression method RMx to the second regression model creation unit 24.

[0057] (5) Creating a second regression model (S5) The creation of the second regression model (S5) will be described with reference to Figures 6 and 7. Figure 6 is a diagram for explaining an example of the second regression model creation process by the second regression model creation unit 24.

[0058] In creating this second regression model (S5), the second regression model creation unit 24 acquires, from the regression method group 33, the regression method RMx corresponding to the identification information 232a of the regression method RMx output from the regression method output unit 232 of the regression method determination unit 23, as shown in FIG. 7, and creates a second regression model 2M based on all data sets 321 included in the data set 32 ​​using the regression method RMx, as shown in FIG. 6, and outputs it to the physical property value prediction unit 25.

[0059] (6) Prediction of physical properties (S6) In this prediction of physical properties (S6), the physical property prediction unit 25 uses the second regression model 2M output from the second regression model creation unit 24 to predict the physical properties based on the material data of the new polymer composition input from the input unit 4, and the display unit 5 displays the physical properties.

[0060] <Actions and Effects of the Present Embodiment> According to this embodiment, the following actions and effects are achieved. (a) Even if the data set 32 ​​does not have a sufficient number of data and is sparse, the number of times the data set 32 ​​is divided is set according to the blank rate of the data, a first regression model is created for each division for each regression method, and the errors of each first regression model are averaged to evaluate the regression method. This reduces the influence of data bias that occurs when the data is divided, and makes it possible to select a regression method that corresponds to the required physical properties. (b) By selecting the regression method with the highest prediction accuracy from among a plurality of predetermined regression methods and creating a regression model corresponding to the selected regression method using all the data in dataset 32, it is possible to make highly accurate predictions of physical property values ​​corresponding to the required physical properties. [Example]

[0061] The number of data in Dataset 32 ​​was set to 350, the blank rate of Dataset 32 ​​was set to 40%, the number of training data and the number of evaluation data were divided into 7:3, and an experiment was conducted to calculate the average error rate using the Gaussian process regression method.

[0062] As a result, in Comparative Example 1, in which a regression model was created and evaluated by dividing data set 32 ​​once using the hold-out method, the first average error rate was 6%. In Comparative Example 2, in which a regression model was created and evaluated by dividing data set 32 ​​once again using the hold-out method, the first average error rate was 11%. On the other hand, according to the present example, in which data set 32 ​​was divided 30 times, the second average error rate was 2% or less, and it is clear that more accurate prediction of physical property values ​​can be performed than with the hold-out method.

[0063] (Summary of the embodiment) Next, the technical ideas grasped from the above-described embodiments will be described by using the reference numerals and the like in the embodiments. However, the reference numerals in the following description do not limit the components in the claims to the members and the like specifically shown in the embodiments.

[0064] [1] A data generation unit (21) that divides a data set (321) of material data indicating blending information of materials constituting a composite material and physical property data indicating physical property values ​​of the composite material into a data set (321) for each of a plurality of types of composite materials so that the combinations of the data set (321) are different from each other, thereby generating a plurality of training data groups and a plurality of evaluation data groups; a first regression model creation unit (22) that creates a plurality of first regression models for each of a plurality of predetermined regression methods based on the plurality of learning data groups; a regression method determination unit (23) that evaluates the prediction accuracy of the plurality of first regression models for each regression method based on the plurality of evaluation data groups and determines the regression method with the highest evaluation among the plurality of regression methods; a second regression model creation unit (24) that creates a second regression model based on the plurality of data sets (321) included in the data set (32) using the regression method with the highest evaluation; a property value prediction unit (25) for predicting a property value based on input material data using the second regression model; and

[0065] [2] The apparatus for predicting physical properties of a composite material (1) described in [1] above, wherein the regression method determination unit (23) comprises: an error calculation unit (231) that calculates, for each regression method and each first regression model, a first prediction error by averaging errors by the first regression model based on a plurality of evaluation data included in the evaluation data group, and calculates, as the prediction accuracy of the regression method, a second prediction error by averaging the first prediction errors by the plurality of first regression models; and a regression method output unit (232) that compares the prediction accuracy calculated for each regression method and outputs the regression method with the highest evaluation.

[0066] [3] The device (1) for predicting physical properties of a composite material according to [1] or [2], wherein the second regression model creation unit (24) creates the second regression model based on the data set that constitutes the data set (32).

[0067] [4] The device (1) for predicting physical properties of a composite material according to any one of [1] to [3], wherein the number of the data sets (321) constituting the data set (32) is 100 or more and less than 100,000.

[0068] [5] The device (1) for predicting physical properties of a composite material according to any one of [1] to [4], wherein the blank rate of the dataset where no data exists is 30% or more and 80% or less.

[0069] [6] The number of times the data set is divided by the data generation unit is a number corresponding to a blank rate where no data exists in the data set. The apparatus for predicting physical properties of a composite material according to any one of [1] to [5] above.

[0070] [7] The composite material property value prediction device (1) according to any one of [1] to [6], wherein the composite material is a polymer composition containing a plurality of resins and a plurality of compounding agents as the material.

[0071] [8] A computer includes a data generation unit (21) that divides a data set (321) of material data indicating blending information of materials constituting a composite material and physical property data indicating physical property values ​​of the composite material into a data set (321) for each of a plurality of types of composite materials so that the combinations of the data set (321) are different from each other, thereby generating a plurality of training data sets and a plurality of evaluation data sets; a first regression model creation unit (22) that creates a plurality of first regression models for each of a plurality of predetermined regression methods based on the plurality of learning data groups; a regression method determination unit (23) that evaluates the prediction accuracy of the plurality of first regression models for each regression method based on the plurality of evaluation data groups and determines the regression method with the highest evaluation among the plurality of regression methods; a second regression model creation unit (24) that creates a second regression model based on the plurality of data sets (321) included in the data set (32) using the regression method with the highest evaluation; a physical property value prediction unit (25) that predicts a physical property value based on input material data using the second regression model;

[0072] [9] A generation step (S1) of dividing a data set (321) of material data indicating blending information of materials constituting a composite material and physical property data indicating physical property values ​​of the physical properties of the composite material into a data set (32) for each of a plurality of types of composite materials so that the combinations of the data set (321) are different from each other, thereby generating a plurality of training data sets and a plurality of evaluation data sets; a first creation step (S2) of creating a plurality of first regression models for each of a plurality of predetermined regression methods based on the plurality of learning data groups; The prediction accuracy of the plurality of first regression models is evaluated for each of the regression methods based on the plurality of sets of evaluation data, and the regression method (RM) having the highest evaluation among the plurality of regression methods is selected. xa determination step (S4) of determining a second creation step (S5) of creating a second regression model based on a plurality of the data sets (321) included in the data set (32) using the regression method with the highest evaluation; a prediction step (S6) of predicting a physical property value based on input material data using the second regression model.

[0073] Although the embodiments of the present invention have been described above, the embodiments of the present invention are not limited to the above-described embodiments, and various modifications and implementations are possible.

[0074] Furthermore, some of the components of the above-described embodiments may be omitted or modified, and steps may be added, deleted, modified, or replaced in the flow of the above-described embodiments. [Explanation of symbols]

[0075] 1...physical property prediction device, 2...prediction processing unit, 3...storage unit, 4...input unit, 5...display unit, 21...data generation unit, 22...first regression model creation unit, 23...regression method determination unit, 24... second regression model creation unit, 25... physical property prediction unit, 31... physical property prediction program, 32...Dataset, 33...Regression methods, 34...Training dataset, 35...Evaluation dataset, 231...Error calculation unit, 232...Regression method output unit, 232a...Identification information, 320...Composition ID, 321...data set, 321a...material data, 321b...physical property data, LDG1~LDG n ...Training data set, TDG1 to TDG n ...Evaluation data set, LD...training data, TD...evaluation data, RM1~RM m ...regression methods, 1M1~1M n ...First regression model, 1MPE1~1MPE n ...first average error rate, 2MPE1~2MPE m …second mean error rate, 2M…second regression model

Claims

1. a data generation unit that divides a data set, which includes material data indicating blending information of materials constituting a composite material and physical property data indicating physical property values ​​of physical properties related to the composite material, for each of a plurality of types of composite materials a plurality of times so that the combinations of the data sets are different from one another, thereby generating a plurality of training data groups and a plurality of evaluation data groups; a first regression model creation unit that creates a plurality of first regression models for each of a plurality of predetermined regression methods based on the plurality of learning data groups; a regression method determination unit that evaluates the prediction accuracy of the plurality of first regression models for each of the regression methods based on the plurality of evaluation data groups, and determines the regression method with the highest evaluation among the plurality of regression methods; a second regression model creation unit that creates a second regression model based on the plurality of data sets included in the data set using the regression method with the highest evaluation; a physical property value prediction unit that predicts a physical property value based on input material data using the second regression model; A physical property value prediction device for a composite material, comprising: The number of times the data set is divided by the data generation unit is equal to or greater than the numerical value of the blank rate where no data exists in the data set. A device for predicting the physical properties of composite materials.

2. The regression method determination unit includes an error calculation unit that calculates, for each of the regression methods and each of the first regression models, a first prediction error by averaging errors due to the first regression model based on a plurality of evaluation data included in the evaluation data group, and calculates, as a prediction accuracy of the regression method, a second prediction error by averaging the first prediction errors due to the plurality of first regression models; and a regression method output unit that compares the prediction accuracy calculated for each of the regression methods and outputs the regression method with the highest evaluation. The apparatus for predicting physical property values ​​of a composite material according to claim 1 .

3. the second regression model creation unit creates the second regression model based on the data sets that constitute the data set. The apparatus for predicting physical property values ​​of a composite material according to claim 1 or 2.

4. The number of the data sets constituting the data set is 100 or more and less than 100,000. The apparatus for predicting physical property values ​​of a composite material according to claim 1 .

5. The blank rate of the dataset, where no data exists, is 30% or more and 80% or less. The apparatus for predicting physical property values ​​of a composite material according to claim 4.

6. The composite material is a polymer composition containing a plurality of resins and a plurality of compounding agents. The apparatus for predicting physical property values ​​of a composite material according to claim 1 .

7. Computer, a data generation unit that divides a data set, which includes material data indicating blending information of materials constituting a composite material and physical property data indicating physical property values ​​of physical properties related to the composite material, for each of a plurality of types of composite materials a plurality of times so that the combinations of the data sets are different from one another, thereby generating a plurality of training data groups and a plurality of evaluation data groups; a first regression model creation unit that creates a plurality of first regression models for each of a plurality of predetermined regression methods based on the plurality of learning data groups; a regression method determination unit that evaluates the prediction accuracy of the plurality of first regression models for each of the regression methods based on the plurality of evaluation data groups, and determines the regression method with the highest evaluation among the plurality of regression methods; a second regression model creation unit that creates a second regression model based on the plurality of data sets included in the data set using the regression method with the highest evaluation; a physical property value prediction unit that predicts a physical property value based on input material data using the second regression model; A physical property value prediction program for a composite material that functions as The number of times the data set is divided by the data generation unit is equal to or greater than the numerical value of the blank rate where no data exists in the data set. A program for predicting the physical properties of composite materials.

8. A computer comprising: a generation step of dividing a data set, which includes material data indicating blending information of materials constituting a composite material and physical property data indicating physical property values ​​of physical properties related to the composite material, for each of a plurality of types of composite materials a plurality of times so that the combinations of the data sets are different from one another to generate a plurality of training data sets and a plurality of evaluation data sets; a first regression model creation step of creating a plurality of first regression models for each of a plurality of predetermined regression methods based on the plurality of learning data groups; a determination step of evaluating the prediction accuracy of the plurality of first regression models for each of the regression methods based on the plurality of evaluation data groups, and determining the regression method with the highest evaluation among the plurality of regression methods; a second generation step of generating a second regression model based on a plurality of the data sets included in the data set using the best-performing regression method; a prediction step of predicting a physical property value based on input material data using the second regression model; A method for predicting physical properties of a composite material, comprising: The number of times the dataset is divided in the generating step is equal to or greater than the numerical value of the blank rate where no data exists in the dataset. Methods for predicting the physical properties of composite materials.

Images