Screening of Performance Measure Indices for PHM System of Armored Vehicles Based on PCA

Building a comprehensive index system is the premise and basis for the scientific evaluation of product performance, and the index screening is the key in the construction of index systems. To solve the problems of high redundancy and low pertinence of the original index set in the constructing process of product performance index system, the principal component analysis method and the correlation analysis were used to reduce the indices. This paper used 14 user performance indices for example. The principal component analysis method was used to screen them, and the 10 performance indices with the largest principal component load value were chosen. Due to the possibility of overlapping information between the selected indices, the correlation analysis was used to remove two relevant indices. The remaining eight indices were used to replace the original index set. To verify the feasibility and effectiveness of the method, the information contribution rate of the remaining indices was calculated. The results show that the remaining indices can present 94.38% of the information the original index set can do. It can be assumed that this method is feasible and effective.