Model Based On Svm Classification Of The Coal And Gas Outburst Forecast

Coal and gas outburst has been the focus that the international and domestic coal circles have paid attention to all the time. There are many coalmines threat to security risks in China over the years. Especially,The coal and gas outburst most are serious and common in recent years. Because of backward prediction, decision-makers cannot access to information of security risk timely so workers cannot be divorced from the scene before the accident, result in serious loss of life and property. How to establish effective coal and gas outburst prediction model for the coal mine and how to provide accurate and reliable basis for decision-makers, avoid or reduce loss of lives and property for the coal are the coal face of the common issues. So search on coal and gas outburst prediction model for classification of great practical significance.Currently, most of the coal domestic combination of soft measurement and sensor on the classification of coal and gas outburst forecast, as the grade of coal and gas outburst is a multi-category classification problem, this modeling technology directly affect levels of coal and gas outburst forecast accuracy. This in-depth study of the domestic coal and gas outburst forecast modeling technique on the classification, Taking into account the establishment of BP neural network in classification of coal and gas outburst forecast model exists defects with poor promote and predict long time, at the same time the Support Vector Machine can be used to handle multiple classification, so using Support Vector Machines to establish classification of coal and gas outburst prediction model. For SVM has superiority of processing large-scale data and Training speed reduced, also prediction speed? This article proposed one algorithm that called RS-C-SVC based on data set reduction. The algorithm start with set reduction, introduction of C-SVC, solve the problem that finding SV. Delete noise data using RS-C-SVC to reduce data set. Increase the training speed effectively, also prediction speed and classification accuracy. Satisfied the need of coal and gas outburst real time forecast. Experimental results show that the algorithm can effectively improve the prediction accuracy. At the same time the cost of computation can be a big problem to solve, reduce the space occupied by memory, improved operating efficiency to shorten the time, for the coal and gas outburst forecast categories quickly and effectively provides a reliable basis.