Information processing method, machine learning method, information processing device, and machine learning device

Abstract

To provide an information processing method that enables improved prediction performance when predicting product characteristics.SOLUTION: An information processing method includes: a manufacturing condition acquisition step of acquiring manufacturing condition information consisting of one or more manufacturing condition parameters defining manufacturing conditions when manufacturing a product from multiple raw materials; an evaluation acquisition step of acquiring evaluation information consisting of one or more evaluation parameters; an importance acceptance step of accepting importance for the one or more manufacturing condition parameters and the one or more evaluation parameters; and a characteristic acquisition step of setting, based on probabilities each corresponding to the importance of each manufacturing condition parameter and each evaluation parameter, in a learning model, usage variables to be used as explanatory variables and quasi-explanatory variables for the one or more manufacturing condition parameters and the one or more evaluation parameters and unused variables to be not used as explanatory variables and quasi-explanatory variables, and acquiring characteristic parameters by inputting the manufacturing condition information and the evaluation information to the learning model in which the usage variables and the unused variables are set.SELECTED DRAWING: Figure 5

Images