不确定支持向量机在洪水预测模型中的应用  

Application of uncertain support vector machine in flood forecasting model

作　　者：李晓丽[1] 周小健[2] 沈钢纲[1] 何晓芳[1] 

机构地区：[1]南京工业大学电子与信息工程学院,江苏南京210009 [2]兰州理工大学计算机与通信学院,甘肃兰州730050

出　　处：《兰州理工大学学报》2012年第3期107-110,共4页Journal of Lanzhou University of Technology

基　　金：江苏省高校自然科学基金(09KJB520006;11KJD520007);南京大学软件新技术国家重点实验室开放基金(KFKT2008B15);东南大学计算机网络和信息集成教育部重点实验室(K93-9-2010-04);南京工业大学学科预研基金(44209105)的联合资助

摘　　要：准确及时地进行洪水预测对洪水预报、洪水实时调度及水资源的合理调度起着非常关键的作用.提出一种粗糙集理论和支持向量机相结合的洪水预测模型,利用粗糙集理论对支持向量机的输入数据集进行约简预处理,通过发现数据间的关系去掉冗余输入信息,简化输入空间的表达信息,提高支持向量机训练的速度,获得较高的预测精度.实验结果表明,该模型能提高支持向量机训练的速度,获得较高的预测精度.Accurate and timely flood prediction plays a key role in flood forecast, real-time flood control, and water resource rational planning. A flood predicting model was presented with rough set theory combi- ning with support vector machine. The input data set of the support vector machine was pre-processed with reduction on the basis of the rough set theory. By means of finding the relation of input data to each other, the redundant input informations were dropped out and the expressed informations in the input space were reduced, so that the training speed of the support vector machine was improved and a higher prediction accuracy obtained. The experimental result demonstrated that this model coud reduce training time and improve prediction precision.

关 键 词：粗糙集 属性约简 支持向量机 洪水预测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]