2019 年 139 巻 8 号 p. 850-857
Since lots of systems in the industries have non-linearity and those structures are generally complicated, it is difficult to express them as mathematical models. The database-driven modeling (DDM) method which is a kind of Just-In-Time(JIT) modeling has been proposed as a method to construct a non-linear model. However, DDM method can not improve modeling accuracy in a complicated system including many needless variables. This study introduces the variable evaluation/selection method based on a random forest to improve the modeling accuracy of DDM method. The random forest can quantify the degree of contribution for variable prediction as importance. The effectiveness of the proposed scheme is numerically verified by some simulation examples.
J-STAGEがリニューアルされました! https://www.jstage.jst.go.jp/browse/-char/ja/