원문정보
초록
영어
A modeling method which can predict the shelf life of various types of spare parts in a relatively short time is put forward in this article. At present, it is difficult to solve the problem of mass modeling because the shelf life prediction models for different kinds of spare parts are of great diversification. In this paper, the best fitting nonlinear variables are selected by Gram-Schmidt regression method, and the detailed steps of automatic modeling process are given, which have advantages of strong robustness and are easy in programming. Especially, it can eliminate the influence of multicollinearity among alternative models effectively. By using natural rubber heating elongation data, an example is taken to demonstrate the process of automatic modeling. The nonlinear regression models selected by automatic modeling process are consistent in Dakin equation, and the predict values of natural rubber shelf life are included in the storage period given by manufacturing plant.
목차
1. Introduction
2. Key Steps of Automatic Modeling Process
2.1 Similar Linear Regression Model
2.2 Gram-Schmidt Orthogonalization Method
2.3 Gram-Schmidt Regression Method
3. Automatic Modeling Process of Nonlinear Regression Model
4. Case Study
5. Conclusions
References