摘要
以径向基函数作为核函数,利用微粒群(PSO)算法的全局寻优特性进行支持向量机(SVM)的参数辨识.在微粒群搜索参数前,先对参数进行指数变换,使[0,1]和[1,∞]有着相同的搜索概率.微粒群算法的适应值函数是以支持向量机模型的推广能力为标准的,讨论了测试样本的最小误差和留一法对支持向量机学习方法推广能力的两种估计.最后以长江宜昌站的月径流资料为例,分别采用ARMA模型、季节性ARIMA模型、BP神经网络模型以及所建立的支持向量机模型进行模拟预测,结果显示了该模型的有效性.
The global searching performance of particle swarm optimization（PSO）algorithm is applied to the identification of the parameters of support vector machine（SVM）in which radial basis function（RBF）is used as kernel function.Before particle swarm searches the parameters,the parameters are transformed into exponent,which makes interval[0,1]and interval[1,∞]have the same searching probability.Adaptive value function of PSO algorithm takes the generalization capability of SVM model as the criterion;thus two kinds of estimation of the generalization capability of SVM,the minimum error of test samples and leave-one-out method,are discussed.Taking Yichang Station in Yangtze River as an example,ARMA model and seasonal ARIMA model,BP neural network model and the proposed SVM model are used to forecast the monthly discharge,respectively.The simulation and forecasting results show the effectivity of the proposed SVM model.
出处
《大连理工大学学报》
EI
CAS
CSCD
北大核心
2011年第1期115-120,共6页
Journal of Dalian University of Technology
基金
国家自然科学基金资助项目(50909012)
水文水资源与水利工程科学国家重点实验室开放基金资助项目(2009490211)
关键词
径流中长期预报
参数辨识
微粒群算法
支持向量机
mid-and long-term runoff forecasting
parameter identification
particle swarm optimization（PSO）algorithm
support vector machine（SVM）
