Prediction of coal and gas outburst grade based on factor analysis and SVM model

In view of the numerous affecting factors of coal and gas outburst and the complex nonlinear relationship, this paper analyzes the main affecting factors of coal and gas outburst in C9 coal seam of a mine in Jinsha County, Guizhou Province. Combined with factor analysis (FA) method and support vector machine (SVM) algorithm, constructed the FA-SVM coal and gas outburst prediction model. The model uses factor analysis to reduce the attributes of high-dimensional original samples, and obtains three common factors that maintain the correlation characteristics of 82.227% of the original data. The first 15 sets of common factors are used as training sets, and the last 5 sets of data are used as the test set, and input into MATLAB for support vector machine algorithm training and prediction. By comparing the prediction results of BP neural network and SVM model, it is found that the FA-SVM model has the highest prediction accuracy, which is 60% and 40% higher than the prediction accuracy of the above two models, and more suitable for coal and gas outburst prediction.