A kernel independent component analysis (KICA) was used for nonlinear feature extraction. A fault classification model based on LSSVM was proposed.

利用核独立成分分析（KICA）进行非线性特征提取，然后用最小二乘支持向量机建立故障分类模型。