The stability of soft rock roadway supporting and the analysis of fractal feature of surrounding rock crack

The rock burst is one of the major problems in deep mining. How to predict the rock burst effectively and to reduce the disaster caused by the rock burst, has great significance. The rock burst is affected by many complex factors, so the forecast of rock burst intensity is a nonlinear, high dimensional, multiclass pattern recognition problem with small samples. A support vector machine is suitable for solving this pattern recognition problem. In this paper, the principle of support vector machine was introduced, the main influence factors of rock burst were given, and a new forecast method of rock burst intensity based on one-against-one SVM classification was presented, through learning the small training samples collected from a mine, the complicated nonlinear mapping relationship between degree of rock burst and its affected factors was established by the proposed method. The case study shows that the method is feasible, easy to be implemented. So the proposed method is very attractive for a wide application in forecasting rock burst.