Study on the selection of predication parameters on mining subsidence based on support vector machine

In order to establish selection model of mining subsidence predicting parameters,which has self learning ability and with high accuracy.In this paper,using principal component analysis preprocessing the data in the literature,we have established the prediction parameters of mining subsidence selection model using support vector machine,based on radial basis function (RBF),by selecting main components factor with cumulative variance reaches 96.79% of 6 and surface subsidence factor as the input and output variables.Results show under the circumstances of less training samples Support vector machine (SVM) model,has high precision and strong generalization ability,the prediction accuracy and prediction stability is better.which was proved contrasting average relative error and root mean square error.