Experimental code based on MATLAB and PYTHON

The experimental code consists data processing and function realization, (1) Data processing. The code of state parameters processing is used for the selection of state parameters standardization of sample processing. The code of label initialization for regression is applied to generate a quantitative degradation value for supervised fine-tuning for the model of regression analysis. The code of PCA use the principal component analysis method to preliminarily evaluate the samples, and roughly distinguish the samples according to the eigenvalues, so as to obtain enough sample labels for reverse fine-tuning of deep learning model. (2) Function realization. The code of regression analysis contains the models of DBN and BP and their regression analysis results. The code of fault classification is used to compare the diagnostic accuracy of DBN, BP and SVM. The code of time series prediction includes the model of LSTM, RNN and GUR. All the related prediction and classification results are also given in the folder.