Research on Gas Outburst Prediction in Coal Mining in China Based on a New Kind of SVM Algorithm

Abstract

The research on predicting gas outburst hazard on the basis of the unbalanced problem of gas monitoring samples under usual circumstances is given in this paper. Combined with outburst-preventing monitoring parameters, a new kind of method of predicting gas outburst hazard based on v-SVM algorithm through analyzing features of real-time gas monitoring data and extracting parameters of gas concentration real-time variation trend, parameters of gas variation rates and feature parameters of gas emission is put forward in this paper. And the application shows that this algorithm for the prediction of actual gas emission in coal mines is effective.