A SVM and model tree based regression model and its application in predicting the amount of gas emitted from coalface

作者：Li Chaoqun; Li Hongwei^*

来源：Journal of Basic Science and Engineering, 2011, 19(3): 370-378.

DOI：10.3969/j.issn.1005-0930.2011.03.003

• Coal mine production
• Combined model
• Mine gas
• Mining process
• Model tree
• Model trees
• Prediction methods
• Prediction model
• Regression model
• Typical application

The mine gas is one of the most unsafe factors in coal mine production. Predicting gas emission directly affects the economic and technological indicators of the mining process. From the data mining and machine learning point of view, it is a typical application of regression analysis. Support vector machine(SVM)and model tree have already demonstrated a superior performance among different regression models. This paper applies them to build prediction models for the amount of gas emitted from coalface. The experimental results show that their precisions suffice the practical use, and both are feasible and reasonable prediction methods. Inspired by their success, we single out a combined regression model based on support vector machine and model tree, and use it to predict the gas emission. The experiments show that the combined model significantly outperforms single regression models.

• 出版日期2011