摘要
为了克服支持向量机参数人工选择的盲目性和依靠经验的缺陷,采用遗传算法优化支持向量机的C,g两个参数,实现支持向量机参数的优选,减少参数选择的工作量,并提高模型的预测精度。并将遗传算法优化的支持向量机模型应用到浑河流域干旱预测中,通过matlab编程建立该模型。以海城、大洼、辽阳、沈阳4个站的降雨量进行建模预测,结果表明:模型的预测值和真实值拟合度较高,模型的预测精度满足要求,说明该模型用于浑河流域的干旱预测具有可行性,为决策部门能及时有效地指导和部署抗旱工作、合理利用分配现有的水资源提供依据。
In order to overcome the defects on the artificial choice blindness for support vector machine parameters and dependence on the experience, we used genetic algorithm to optimize the support vector machine C, g parameters, realizing the optimization of the parameters of support vector machine, reducing the workload of parameter choice and improving the prediction precision. Genetic algorithm support vector machine model was applied to the flood drought prediction of Hun River Basin, through the MATLAB programming to build the model. In Haieheng, Dawa, Liaoyang and Shenyang, the rainfall prediction model was built. The model predictive value and real value was well fitted, the prediction precision met the requirements. So the model for flood drought of Hun River basin prediction is feasible, useful for decision-making department, deployment of drought fighting work and reasonable utilization and allocation of water resources.
出处
《沈阳农业大学学报》
CAS
CSCD
北大核心
2013年第2期190-194,共5页
Journal of Shenyang Agricultural University
基金
高等学校博士学科点专项科研基金联合资助项目(201112103110007)
沈阳市科技局农业科技攻关项目(F12-129-3-00)
关键词
遗传算法
支持向量机
干旱预测
降雨量
genetic algorithm
support vector machine
droughts prediction
rainfall.
