This paper proposes a supervised machine learning model based on a decision tree (DT), multilayer perceptron (MLP), and ranking support vector machine (Rank-SVM) for the diagnosis of resistance variable multifault location (RVMFL) in a mine ventilation system. The feasibility of the method and the predictive performance and generalization ability of the model were verified using a tenfold cross-validation of a multifault sample set of a 10-branch T-shaped angle-joint ventilation network and a 54-branch experimental ventilation network. The results show that the three models, DT, MLP, and Rank-SVM, can be used for the diagnosis of RVMFL in mine ventilation systems, and the prediction performance and generalization ability of the MLP and DT models perform better than the Rank-SVM model.
Related Posts
