Mechanical equipment fault identification based on machine learning algorithm under unbalanced data

Aiming at the fault diagnosis of mechanical equipment under category imbalance, a combined model XGBoost-SMOTEENN-Decision Tree was constructed to identify mechanical equipment faults. Firstly, SMOTEENN algorithm is used to deal with the unbalanced distribution of “normal operation” and “fault” samples in mechanical equipment data; Secondly, the Decision Tree algorithm is used to evaluate the importance of features. Then we compare the three machine learning models XGBoost, AdaBoost and Random Forest by accuracy, precision, recall and F1 value model performance evaluation index, and find that XGBoost model has more significant classification ability, stronger generalization ability, and is more suitable for mechanical equipment fault diagnosis. Finally, the optimal classification model XGBoost is used to compare the classification effect before and after data balancing treatment, and it is found that the fault identification effect of mechanical equipment is better after data balancing treatment.