Reliability Evaluation of Rolling Bearings Based on Stochastic Proximity Embedding and Logistic Regression
A rolling bearing operation reliability evaluation method based on stochastic adjacent embedding and logistic regression model was proposed,which could solve the reliability problems in rolling bearing operation.Firstly,feature extraction of vibration signals is carried out to construct feature parameter sets;secondly,the random adjacent embedding dimensionality reduction algo-rithm is used to reduce the dimensionality of the feature parameter set to form the low-dimensional feature vector set;thirdly,the wavelet denoising method is used to de-noise the feature vector set;finally,the denoised data were put into the logistic regression model to evaluate the reliability of rolling bearings,and the validity of the proposed method was proved by the rolling bearing vi-bration signal dataset from Xi'an Jiaotong University.
Rolling BearingStochastic Proximity EmbeddingLogistic RegressionWavelet DenoisingReliability EvaluationFeature Extraction
万方数据
维普
