基于RS-SVM模型的煤与瓦斯突出多因素风险评价  

Study on Multiple Factors Risk Evaluation of Coal and Gas Outburst Based on RS-SVM Model

作　　者：刘俊娥[1] 曾凡雷[2] 郭章林[3] 

机构地区：[1]北京物资学院信息学院 [2]河北工程大学经济管理学院,河北邯郸056038 [3]华北科技学院土木工程系

出　　处：《中国安全科学学报》2011年第7期21-26,共6页China Safety Science Journal

基　　金：国家"十一五"科技支持计划(2007BAB18B01);北京市属高等学校人才强教计划资助项目(PHR200906210)

摘　　要：为挖掘瓦斯突出风险与煤矿开采中各影响因素间的关系,应用支持向量机(SVM)理论从模式判别角度分析瓦斯突出风险与各地质因素组成的特征向量间的判别关系,基于粗糙集(RS)理论对待分析数据进行知识约简,提取核心判别指标,建立基于粗糙集-支持向量机(RS-SVM)的瓦斯突出风险判别模型。研究结果表明,RS知识约简方法可以很好地对原始数据中的冗余指标进行约简,通过对约简后指标数据进行SVM回归分析,可对煤与瓦斯突出模式进行很好的判别,所建立的瓦斯突出风险判别模型较一般SVM模型具有更高的预测精度,同时指标约简过程降低SVM运算中的复杂度,提高运算效率。In order to study the relationship between gas outburst risk and its affecting factors in coal mine,with the application of SVM theory,the judgment relation between gas outburst risk and the eigenvectors consisted of geological factors is analyzed from the perspective of pattern recognition,then RS model is used to do the knowledge reduction about the original data and to extract the key indexes,so as to establish a risk determination model based on RS-SVM theory.Results show that redundancy of the indicators in original data can be well eliminated with the RS knowledge reduction method,through the SVM regression analysis about data of key indexes,there is a good discrimination model about the coal and gas outburst,and the accuracy of this risk determination model is better than that of the common SVM prediction method.Meanwhile,index reduction process reduces the calculation complexity of the SVM operations and improves its operational efficiency.

关 键 词：煤与瓦斯突出 粗糙集(RS)理论 支持向量机(SVM) 风险评价 指标约简 

分 类 号：X936[环境科学与工程—安全科学]