The Application of Improved Local Preserving Projection Dimension Reduction Method in Spare Parts Classification

Abstract

In order to classify spare parts by machine learning method, it is necessary to reduce the dimension of spare parts because of the large number of characteristic indexes. An improved dimensionality reduction method for local preserving projection is presented in this paper. The model optimizes the parameters in local preserving projection dimensionality reduction by using kernel function parameter estimation method. The accuracy and efficiency of dimension reduction classification model are improved. Finally, through an example analysis, the simulation results show that the proposed dimension reduction method can solve the problem of spare parts classification better. It also improves the accuracy and efficiency of traditional classification methods.