基于RS-SVM的瓦斯涌出量预测研究

Study on Prediction of Gas Emission Based on RS-SVM

针对回采工作面瓦斯涌出量预测中现存的问题,将粗糙集(RS)理论的属性约简功能和支持向量机(SVM)的非线性预测方法相结合建立预测模型。基于现有文献研究,选取13个影响因素作为瓦斯涌出量的自变量。分别将粗糙集约简前的13个影响因素和粗糙集约简后的4个影响因素构成的样本数据结合支持向量机建立SVM预测模型和RS-SVM预测模型。实验表明基于RS-SVM预测模型能够将预测精度由SVM模型的95%增加到99%。解决了工作面瓦斯涌出量预测中现存的问题,提高了预测精度,是一种有效可行的预测方法。

In view of the existing problems in the prediction of gas emission in the working face,a prediction model is established by combining the attribute reduction function of rough set（RS） theory and the nonlinear prediction method of support vector（SVM）.Based on the existing literature research,13 influencing factors were selected as the independent variables of gas emission.The SVM prediction model and the RS-SVM prediction model were established by combining the 13 influencing factors and the four influencing factors after the rough set reduction.Experiments show that the prediction accuracy can be increased from 95% to 99% of the SVM model based on the RS-SVM prediction model.It solves the existing problems in the prediction of gas emission in the working face and improves the prediction accuracy,which is an effective and feasible forecasting method.