Application of projection pursuit method based on accelerated genetic algorithm for evaluation model

Abstract

In view of the evaluation decision-making problems of multi-index system such as supplier comprehensive evaluation, this article adopts the method of dimension reduction, namely using projection pursuit method based on accelerating genetic algorithm optimization. The algorithm can optimize the projection vector, let the high dimensional data through a certain combination of projection into the low dimensional subspace, and obtain the score of the sample from the projection function, so as to make scientific decision. This model avoids the interference of subjective factors to a large extent, and the evaluation results are more accurate and reliable, which provides a new method and thought for other evaluation and decision problems.