The Fault Diagnosis of Rolling Bearing Based on Improved Deep Forest

Rolling bearing fault diagnosis is a meaningful and challenging task. Most methods first extract statistical features then carry out diagnosis. At present, the technology of intelligent identification mostly relies on deep neural network, which has high requirements for computer equipment great effort in hyperparameter tuning. To address these issues, rolling method based improved forest algorithm proposed. Firstly, feature information extracted through multigrained scanning, carried by cascade forest. Considering fitting quality diversity classifier, classifier strategy are updated. In order to verify effectiveness proposed method, comparison made with traditional machine learning method. The results suggest that can identify different types faults more accurately robustly. same time, it very few hyperparameters low hardware.