The PSO-LSSVM model for predicting the failure depth of coal seam floor

Analyzed the samples of the failure depth of coal seam floor collected in mining fields, studied the main influence factors being associated with the failure depth. In order to avoid overfitting problem of artificial neural network (ANN), a new least squares support vector machines (LS-SVM) model is presented to forecast the nonlinear failure depth of coal seam floor under the influence of mining based on particle swarm optimization(PSO) method. PSO is used to choose the parameters of LS-SVM, which can avoid the man-made blindness and enhance the efficiency, even improve the generalization performance. The experimental results show the method is feasible and precise, with reliable theoretical foundation and good practical performance.