Abstract：The purpose of the study on the accuracy and reliability of coal and gas outburst prediction is to improve the quality of prediction and discrimination. By analyzing and comparing, the perturbation learning samples that affect the accuracy of the model are removed. After the reduction of the principal component in the five key attributions that determine coal and gas outburst, The SVM regression model for predicting coal and gas outburst was constructed, and 0.9 was selected as the main component information retention rate, so that the information of 90 % the original variable was retained in the main component; The model parameters are determined by introducing relaxation variables and transforming the SVM convex optimization problem into a linear programming problem. The accuracy and validity of the model in the prediction of coal and gas outburst were verified by the training of 14 samples and the results of 8 samples. The prediction effect is good, and the model algorithm is easy to realize in MATLAB programming, which provides a new idea for coal and gas outburst prediction.
2020, Vol. 30 