Research In Coal And Gas Outburst Prediction Method Based On Optimized Support Vector Machine

Coal and gas outburst has sudden strong and destructive, it is the natural disasters in coal mine.Coal and gas outburst is a complex nonlinear problem which is affected by multiple factors. Statistical learning theory propose Support Vector Machine that it is a new machine learning method,it has strong generalization ability,and it is showed superiority in solving the nonlinear problem of small sample. Therefore, this article, proposed classification model to predict coal and gas outburst danger by using support vector machine (SVM) which is based on its advantages.The penalty parameter C and the kernel function parameter ? of Support Vector Machine influence its generalization and classification prediction accuracy. When we set up the gas outburst prediction of support vector machine model, using the combination of grey relational analysis and entropy method to select the main factors of affecting gas outburst,thus indicator system of classification prediction model is established.The article considered parameters selection as combinatorial optimization problem, set up objective function,applied PSO algorithm and GA to SVM predictive model respectively,so as to realize automatic selection and optimization of the parameters.The measured data of mine production sets as training and testing samples used in gas outburst of PSO-SVM and GA-SVM classification prediction model,simulation results showed that two kinds of optimization algorithm could select the SVM coal and gas outburst model parameters effectively through establishing prediction model. In comparison with GA, PSO showed better learning ability, model calculations consume a very short time and could get better predictive accuracy, it has a good generalization performance.The prediction model of coal and gas outburst based on PSO-SVM Machine in ensuring the coal mine production safety is extremely important.