基于支持向量机的矿压监测数据预测研究  

Prediction of mine pressure monitoring data based on support vector machine

作　　者：孙刚[1] 周华平 孙克雷[2] 

机构地区：[1]阜阳师范学院计算机与信息学院,安徽阜阳236037 [2]安徽理工大学计算机科学与工程学院,安徽淮南232001

出　　处：《阜阳师范学院学报（自然科学版）》2013年第3期46-49,共4页Journal of Fuyang Normal University(Natural Science)

基　　金：国家自然科学基金和神华集团有限责任公司联合基金项目(51174257/E0422);安徽省高校科学研究项目(KJ2013B196)资助

摘　　要：冲击矿压具有非常复杂的行为特征,矿压监测数据隐含着冲击矿压系统参与演化的全部变量的信息,可以利用矿压监测数据对冲击矿压预测。文章首先对矿压监测数据进行混沌分析,构建混沌矿压监测数据的相空间,利用支持向量机建立矿压监测数据的预测模型。经与神经网络预测模型进行比较,文章所提出的基于支持向量机的矿压监测数据预测模型更加准确有效。Rock burst has very complex behavioral characteristics and rock pressure monitoring data implies information of all variables involved in the evolution of rock burst system; therefore rock burst can been predicted by using the mine pressure monito- ring data. Firstly, chaos of mine pressure monitoring data was analyzed in this paper, and then phase space of the chaotic mine pressure monitoring data was constructed. Finally, prediction model of mine pressure monitoring data was built by using support vec- tor machine. Compared with neural network prediction model, the prediction model based on support vector machine proposed in this paper is more accurate and effective for prediction of mine pressure monitoring data.

关 键 词：矿压 混沌分析 支持向量机 预测 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]