基于粒子群优化支持向量机的矿井涌水量预测  

Prediction of Water Inrush in Mine Based on Particle Swarm Optimization-based Support Vector Machine

作　　者：臧大进[1,2] 刘增良[1] 曹云峰[3] 

机构地区：[1]铜陵学院电气工程系,安徽铜陵244000 [2]中国矿业大学,江苏徐州221008 [3]上海交通大学电子信息与电气工程学院,上海200240

出　　处：《凯里学院学报》2010年第6期26-29,共4页Journal of Kaili University

摘　　要：矿井涌水量预测是一项复杂而有难度的技术,受到很多因素的影响.提出基于粒子群优化支持向量机(PSO-SVM)的矿井涌水量预测方法,即将粒子群优化算法(PSO)用于SVM参数优化.它不仅具有很强的全局搜索能力,而且容易实现.经实验结果证明,PSO-SVM的预测输出与实测数据基本一致,其预测精度高于普通的SVM,所有的预测误差都远小于5%的工程许可误差.Prediction of water inrush in mine is a complex and difficult technology, because it is related with many factors. The proposed PSO - SVM method was applied to predict the water inrush in mine in the paper, among which particle swarm optimization （PSO） was used to de- termine free parameters of support vector machine. The method not only had strong global search capability, but also was very easy to implement. Prediction of water inrush in mine examples were used to illustrate the performance of proposed PSO - SVM method. The experi- mental results indicated that the PSO - SVM method can achieve the nearly same result as measured data and higher diagnostic accuracy than normal SVM consequently, whichwas far less than 5% of the project license error.

关 键 词：粒子群优化支持向量机 粒子群优化算法 支持向量机 矿井涌水量 预测 

分 类 号：TD742.1[矿业工程—矿井通风与安全]