因子分析与支持向量机相结合的煤炭自燃预测  

Coal spontaneous combustion prediction model of support vector machine combined with factor analysis

作　　者：邵良杉[1] 马寒[1] 温廷新[1] 

机构地区：[1]辽宁工程技术大学系统工程研究所,辽宁葫芦岛125000

出　　处：《辽宁工程技术大学学报（自然科学版）》2014年第4期433-436,共4页Journal of Liaoning Technical University (Natural Science)

基　　金：国家自然科学基金资助项目(70971059);辽宁省科学研究计划基金资助项目(2010230004)

摘　　要：针对煤炭自然发火预测问题,将因子分析理论与支持向量机算法相结合,构建了煤炭自燃的预测模型.采用SPSS软件对所选取预测煤炭自燃的8个指标进行因子分析,提取出4个主要因子,利用Matlab实现支持向量机模式识别,通过实际数据对模型训练与检验.研究结果表明:经因子分析降维后的SVM预测准确率高于未经因子分析的SVM预测结果,并验证了因子分析与支持向量机相结合可以较高精度实现对煤炭自燃的预测.In order to complete the prediction of coal spontaneous combustion, a prediction model was built based on factor analysis and support vector machine. SPSS software was used to extract 4 main factors from 8 related variables, and Matlab was applied to implement support vector machine for pattern recognition. Training and testing were carried out using actual data. The result shows that its prediction accuracy is higher than the one without factor analysis, and the model based on factor analysis and support vector machine can complete the prediction of coal spontaneous combustion with a high accuracy.

关 键 词：煤炭 自燃 因子分析 支持向量机 数据降维 因子值 交叉验证 预测精度 

分 类 号：X936[环境科学与工程—安全科学]