PROJECT TITLE :

Wavelet Kernel Local Fisher Discriminant Analysis With Particle Swarm Optimization Algorithm for Bearing Defect Classification

ABSTRACT:

Feature extraction and dimensionality reduction (DR) are necessary and useful preprocessing steps for bearing defect classification. Linear native Fisher discriminant analysis (LFDA) has recently been developed as a well-liked method for feature extraction and DR. But, the linear technique tends to convey undesired results if the samples between classes are nonlinearly separated within the input area. To enhance the performance of LFDA in bearing defect classification, a new feature extraction and DR algorithm based on wavelet kernel LFDA (WKLFDA) is presented during this paper. Herein, a new wavelet kernel operate is proposed to construct the kernel operate of LFDA. To ask for the optimal parameters for WKLFDA, particle swarm optimization (PSO) is used; as a result, a new PSO-WKLFDA algorithm is proposed. The experimental results for the artificial information and measured vibration bearing data show that the proposed WKLFDA and PSO-WKLFDA outperform alternative state-of-the-art algorithms.