Research on electric vehicle fault prediction method based on Bagging integrated learning

Article ID: 4281

10 Views

Keywords: Bs_Bagging; CLightGBM; Integrated Learning; Electric Vehicle Fault Prediction; Robustness

Abstract

To address the problems of poor classification performance and low fault detection rate of machine models caused by the imbalance of electric vehicle fault data samples, this paper proposes a Bagging integrated electric vehicle fault prediction model with LightBDM as the base learner improvement based on the BS_Bagging-cLightGBM model. First, the training set is resampled using the Borderline_ SMOTE method in Bagging integrated learning to improve the degree of data imbalance in the training subset and avoid the missing information of small class samples; then, the weight coefficients and regularization terms are embedded in the loss function of the LightGBM base learner to improve the misclassification cost of small class samples in training; finally, we was evaluated and validated, and the experimental results showed that the BS_Bagging-cLightGBM model outperformed the single model in terms of accuracy, recall and F1-score metrics, and showed better prediction capability. The results of the study can provide an important reference for the repair and maintenance of electric vehicles.