Application Of Principal Component-AHP Combination Method In Performance Comprehensive Evaluation

Principal component analysis is a very important method in the comprehensive evaluation,which can effectively empower the index,so it has a wide range of applications in the comprehensive evaluation.The principal component analysis method is objective weighting method,which has the function of reducing dimension and can eliminate the interference of man and has better objectivity.However,there is a certain drawback when using only the statistical method of principal component analysis.In the case of empowerment,a simple objective weighting law makes the results inconsistent with reality.So,the paper use analytic hierarchy process,through the subjective weight of the way to adjust the objective weight.The innovation of this paper is to combine AHP,principal component analysis and cluster analysis.First,we use the principal component analysis to get a weight index to remove human factors.In the past research work,the principal component analysis mainly has the function of reducing dimension.Principal component analysis which can effectively remove the influence of human factors.Not only makes the objectivity of the weight to be reflected,but also can retain most of the information of the information.However,since the principal component analysis is completely dependent on the information of the data itself,it has some limitations,which leads to the deviation between the weight determined by the principal component and the actual importance of the indexes.The thesis adopts the analytic hierarchy process,carries on the subjective empowerment by the expert scoring,and corrects the objective weighting result.This is because the expert score can reflect the expert's expertise,work experience,etc.,contribute to the data to make more realistic evaluation,so you can use the subjective weight to modify the weight.The paper use principal component analysis method to decide the weight of the indexes.According to the contribution rate of the principal component to calculate the weight.And then,use the analytic hierarchy process,according to experts score to adjust and correct the weight of principal component analysis.To obtain a set of subjective empowerment and objective weight of one of the more practical indicators of comprehensive weights.Finally,the clustering method is used to cluster the samples.The clustering results are consistent with the times of sample selection,and the feasibility of the method of principal component-level analysis is verified.