Abstract:Aiming at the shortcomings of existing SVM classification algorithms in high-dimensional small sample fault feature classification,fitness function selection and core parameter optimization,a mechanical fault classification algorithm based on SOA-SVM was proposed.The wavelet threshold function was used to denoise the original fault signal,then several random behaviors of the crowd were simulated based on SOA algorithm,and the optimal moving direction and step size of the fault data points were selected.Finally,the optimal geometric distance from the hyperplane of SVM classifier was found to improve the performance of fault data classification of classical SVM classifier.The simulation results show that the proposed fault classification algorithm has stronger parameter optimization performance,and higher classification accuracy can be obtained in the classification of multiple high-dimensional small sample data sets.