Comprehensive Evaluation Method Based on Principal Component Analysis and Class Coverage Discriminant Analysis

Aiming at the problem of multi-industry comprehensive evaluation, this paper discusses a comprehensive evaluation method based on principal component and class coverage discriminant analysis. This method adopts principal component analysis and the standard value of industry performance to classify the sample data, and the original sample data are divided into five levels, namely, excellent, good, average, poor, and very poor, and a comprehensive evaluation model is constructed by using the idea of class coverage and nearest neighbor principle. To verify the effect of the model, the paper collected the relevant data of 108 listed companies in 4 industries in 2022. The sample data were divided into a training group and a test group, three schemes were set up, and the comprehensive evaluation model was established by using the methods discussed in the paper, respectively. Then, a comparative analysis was conducted with Fisher discriminant analysis, k-nearest neighbor discriminant analysis, and nonlinear discriminant analysis method based on the cover class problem. The experimental results show that compared to the common supervised classification comprehensive evaluation method, the proposed comprehensive evaluation method has better stability and classification accuracy.