改进的模糊多类支持向量机算法在瓦斯预警中的应用  

Gas Early Warning Based on Improved Algorithms of Fuzzy Support Vector Machines

作　　者：宋文津[1] 

机构地区：[1]承德石油高等专科学校,河北承德067000

出　　处：《煤矿机械》2013年第3期232-233,共2页Coal Mine Machinery

基　　金：河北省科技攻关课题(4213571)

摘　　要：煤与瓦斯危险性的准确预测一直是矿山安全领域的关键技术难题和重大研究课题。支持向量机是在瓦斯预警中广泛使用的一种技术,以统计学习理论和支持向量机为基础,通过研究基于模糊支持向量机的多类分类方法,对原算法进行改进,采用模糊多类支持向量机,并构造模糊隶属函数,同时使用序列最小最优化算法进行求解,以期提高算法的精度和速度。Coal and gas risk accurately predict the field of mine safety has been the key technical problems and major research topic. SVM is widely used in gas warning. This paper through research based on fuzzy support vector machine multi-class classification, an improved algorithm by introducing the fuzzy membership of data samples, using sequential minimal optimization to solve fuzzy multicategory support vector machine, in order to improve the algorithm＇saccuracy and speed.

关 键 词：支持向量机 瓦斯预警 模糊隶属函数 

分 类 号：TD712[矿业工程—矿井通风与安全]