Fault Diagnosis of Rolling Bearings Based on Neural Networks and Decision Trees

Based on the excellent accuracy of Convolutional Neural Network (CNN) models but their long algorithm runtime, and the high efficiency but relatively lower accuracy of decision tree models, this paper proposes a CNDT (CNN combined with Decision Tree) model by integrating the advantages of both approaches. Firstly, the powerful feature extraction capability of CNN is utilized to extract hidden features from the raw data. Then, these extracted features are fed into the decision tree model for outputting the final results. In this study, the CNDT model is validated using a dataset of rolling bearing faults. The results demonstrate that the proposed model achieves higher accuracy and reduces algorithm runtime, providing a feasible solution for rolling bearing fault diagnosis. Future research can focus on further optimizing the performance of the CNDT model and applying it to fault diagnosis tasks in other domains.