基于蚁群算法优化SVM的瓦斯涌出量预测  

Forecasting of Gas Emission Quantity Based on Ant Colony Optimization SVM

作　　者：王晓路[1] 

机构地区：[1]西安科技大学通信与信息工程学院,西安710054

出　　处：《煤炭技术》2011年第5期81-83,共3页Coal Technology

基　　金：电子信息产业发展基金招标项目(XDJ2-0514-27);国家高技术研究发展计划(863计划)(2005AA133070);陕西省教育厅项目(09JC05)

摘　　要：为了提高瓦斯涌出量预测的精度和预测模型的泛化能力,提出了一种基于蚁群算法(ACO)优化支持向量机(SVM)参数的瓦斯涌出量预测方法。在SVM所建立预测模型中各个参数的取值区间内,采用蚁群优化算法计算预测模型各个参数的最佳值,基于最佳参数的SVM建立瓦斯涌出量预测模型。结果表明:采用未优化的SVM建立的预测方法,其个别预测误差相对较大,最大误差为8.11%,平均误差为4.68%,采用ACO对于预测模型的参数进行优化后,预测性能有显著提高,最大误差为4.37%,平均误差为2.89%,表明所建议的方法是有效、可行的。To accurately predict the gas emission quantity and improve forecasting performance,an approach based on the parameters of Support Vector Machine（SVM） being optimized by Ant Colony Optimization（ACO） arithmetic is proposed.ACO is introduced to calculate the best parameters in their respective value interval of SVM forecasting model.The forecasting model of gas emission quantity is established based on the SVM with the best parameters.The results show that some prediction deviations are relatively large with the averaged biases of 4.68% and the maximum error of 8.11% by using the SVM without optimization.The predictable precision is obviously improved to obtain the averaged biases of 2.89% and the maximum error of 4.37% by using the SVM based forecaster based on the best parameters calculated by ACO.It is indicated that the suggested approach is feasible and effective.

关 键 词：瓦斯涌出量 蚁群优化算法 支持向量机 预测 

分 类 号：TD713[矿业工程—矿井通风与安全]