Application of Support Vector Machines in Coal and Gas Outburst Area Prediction

In predicting coal and gas outburst, because of coal and gas outburst insufficient samples, the knowledge-based method in coal and gas outburst prediction were restricted to some extent. The coal and gas outburst prediction model by support vector machines, which had a strong ability to identify the characteristics in a few sample of cases, was put forward to solve the problem in the paper. Factors affecting area of coal and gas outburst were withdrawn as characteristic vectors based on the genetic algorithm according to the natural conditions and the characteristics of the geologic structure. The forecast model of support vector machines was validated with the practical example. The comparison result from support vector machines forecasting and the traditional methods indicated that this SVM (support vector machines, brief named SVM) method could meet the requirement of coal and gas outburst area forecast. The study results proved the validity of the model, and laid foundation for the area forecast of the coal and gas outburst based on support vector machines.