Application Of GA-BP Weighted Network And SVM In Coal And Gas Outburst Prediction

With the rapid development of China’s economy, the demand for coal has increased year by year, coalis an important raw material in metallurgy, chemical industry. Because a large number of mining coal mine accidents have also brought occurs frequently, where coal and gas outburst accident caused by accidents accounted for the largest proportion of all mine. Prominent reason there are a lot of accidents occur frequently, can not accurately predict the highlight of the important reasons is caused by the accident. According to the forecast results can take appropriate safety measures to reduce the risk to a minimum. Therefore, a high accuracy of the prediction model can effectively reduce the accident rate, improve the safety factor.In this paper, the use of neural network and SVM to construct two kinds of outburst prediction models. In the neural network prediction model, using a GA algorithm to optimize the weights and thresholds of neural network, improve network performance; in SVM forecasting model, a reasonable choice of kernel function, and then determine the parameters to be optimized according to the kernel function, compared to each different parameter optimization method to choose the best parameter values of c and g values, which can improve the accuracy of the constructed SVM training and testing.C refers to the penalty factor, G is a parameter setting function.In order to enhance the accuracy of the prediction model before making neural networks and SVM training and prediction, the first of the training data and prediction weights, sizes weights by using gray relational analysis method and seek treatment on the proportion of outstanding this five main factors identified five important factors are:gas pressure, diffuse initial velocity, depth of mining, coal damage type, consistent coefficient. Were right for each value processing factors contribute to the strong influence of factors inhibiting influence of weak factors, so as to improve the prediction accuracy of the results.The simulation results show that:After the GA-BP network and the data weighted SVM prediction model can post a more accurate predict after a prominent coal and gas, has a certain practicality.