Coal and gas outburst prediction model based on principal component analysis and improved support vector machine

Abstract

In order to predict the coal outburst risk quickly and accurately, a PCA-FA-SVM based coal and gas outburst risk prediction model was designed. Principal component analysis (PCA) was used to pre-process the original data samples, extract the principal components of the samples, use firefly algorithm (FA) to improve the support vector machine model, and compare and analyze the prediction results of PCA-FA-SVM model with BP model, FA-SVM model, FA-BP model and SVM model. Accuracy rate, recall rate, Macro-F1 and model prediction time were used as evaluation indexes. The results show that: Principal component analysis improves the prediction efficiency and accuracy of FA-SVM model. The accuracy rate of PCA-FA-SVM model predicting coal and gas outburst risk is 0.962, recall rate is 0.955, Macro-F1 is 0.957, and model prediction time is 0.312s. Compared with other models, The comprehensive performance of PCA-FA-SVM model is better.