Coal Thickness Prediction Based on Support Vector Machine Regression
Zhengwei Li, Xia Shixiong, Niuqiang, Xia Zhanguo · 2007
A novel method based on support vector machine for coal thickness prediction through seismic attribute technology is proposed in this paper. Based on SVM which embodies the Structural Risk Minimization principle, the proposed method is more generalized in performance and accurate than artificial neural network which embodies the Embodies Risk Minimization principle. In order to improve prediction accuracy, grid search and cross-validation are integrated in this paper to select SVM parameter. Error analysis of predicting coal thickness is carried out to prove that SVM could achieve greater accuracy than the BP neural network.