基于GRA-SVM方法的煤与瓦斯突出预测模型研究  

Research on prediction model for coal and gas outburst based on GRA-SVM

作　　者：曲方[1] 安文超[1] 李迎业[2] 李忠群[2] 

机构地区：[1]中国计量学院质量与安全工程学院,浙江省杭州市310018 [2]山西潞安环保能源开发股份有限公司五阳煤矿,山西省长治市046025

出　　处：《中国煤炭》2012年第11期102-106,共5页China Coal

摘　　要：通过分析煤与瓦斯突出的主要影响因素,采取灰熵关联分析法进行关联度排序及特征向量的提取,利用支持向量机强大的模式识别能力,提出基于支持向量机的煤与瓦斯突出预测方法。以典型的煤与瓦斯突出样本为训练实例,利用Matlab平台下的Libsvm工具包建立预测模型,通过测试样本对模型进行验证,表明此模型具有较高的分类精度,适合于解决小样本的突出预测问题。On the basis of the analysis of main influencing factors of coal and gas outburst, the grey relational analysis （GRA） was usedto rank the degree of association and to extract the eigenvectors. By the aid of powerful identifying function of support vector machine （SVM）, a new methodology for predicting coal and gas outburst was proposed. Taking typical cases of coal and gas outburst as training examples, the prediction model was constructed based on Libsvm toolbox of Matlab platform. The verification results showed the prediction model has higher clas- sification accuracy for coat and gas outburst, suitable to solve the prediction problem for small samples of coal and gas outburst in coal mines.

关 键 词：煤与瓦斯突出 灰色关联度 支持向量机 预测 

分 类 号：TD713.2[矿业工程—矿井通风与安全]