Research on anomaly detection of underground equipment based on MS-IF-AdaBoostM1

The complex underground environment in coal mines leads to a high incidence of mechanical and electrical equipment failures, directly impacting production safety and efficiency. Existing anomaly detection methods suffer from insufficient generalization capabilities when applied to multiple scenarios. This paper, based on cumulative frequency rules and duration rules, analyzes production data from a coal mine in Shendong Coal Group and proposes an anomaly detection integrated model（MS-IF-AdaBoostM1） using machine learning technology. This model combines the Isolation Forest model and the AdaBoostM1 model with multi-scale time windows. Experimental results show that the detection accuracy of this model is significantly higher than that of a single model, validating the superiority of the dual-model approach and providing a new method for anomaly detection in underground equipment.