In order to extract the fault features of bearing vibration signals more accurately

weighted fine composite multi-scale spread entropy (wRCMDE) was introduced into bearing fault feature extraction. On this basis

a rolling bearing fault diagnosis method based on wRCMDE and improved Bayesian network was proposed. By calculating wRCMDE of different fault vibration signals and selecting multiple wRCMDE values at appropriate scale as feature vectors

feature samples were formed and input into the Bayesian network optimized by the improved firefly algorithm for fault classification and recognition. Through the analysis of experimental data

the proposed method is compared with the fault feature extraction method based on multiscale dispersal entropy and refined composite multiscale dispersal entropy. Experimental results show that this method can identify the fault types of rolling bearings more accurately

and the recognition rate is higher.