Mining Working Face Time Series Short-term Gas Prediction Based on LS-SVM

At present, one of development direction of mine gas prediction is the statistical learning method. In this paper author firstly introduces the character of SVM, and on this basis give the basic principle of LS-SVM, and at the same time establish LS-SVM regression model. Secondly, the data of time series gas concentration are standardized in the range of [-1, 1], subsequently these data are reconstructed and used for training data and test data. Finally, in the MATLAB7.1 environment, this prediction model is achieved by algorithm procedure. The working face gas outburst data of the 10th coal mine in Hebi is used to train and test this model. According to two examples simulation result shows that this model has well the short-term working face gas predict effects.