A Bayes Discriminant Analysis Method for Predicting the Hazard Classification of Rockburst and its Application

A method to forecast the hazard classification of rockburst by using the Bayes discriminant analysis theory is presented in this paper. The Bayes discriminant analysis (BDA) theory was introduced firstly. Then considering the mining circumstances and geological conditions of rockburst, eight factores reflecting the rockburst, including exploitation depth of coal seam, lithology of stope roof, complicated degree of conformation, dip angel of coal seam, thickness of coal seam, exploitation method, coal pole state and blasting mining/comprehensive mining, were considered and 14 specific indexes were selected to establish a BDA model. 24 samples of the Zhangji Mine in Xuzhou City of China were used as the training and forecasting samples. The prior probability of each collectivity was obtained according to the ratio of training samples and re-substitution method was also introduced to verify the stability of model. Compared with the artificial neural network (ANN) method and support vector machine (SVM) method, the results show that this BDA model has excellent performance, high prediction accuracy and can be used in practical engineering.