基于模糊聚类分析的小波神经网络模型在径流预测中的应用  

Wavelet Neural Networks Based on Fuzzy C-Means Clustering and its Application in the Runoff Forecast

作　　者：徐瑾[1,2] 钟炜[1] 李溪楠[1] 

机构地区：[1]天津理工大学管理学院,天津300384 [2]天津大学管理学部,天津300072

出　　处：《数学的实践与认识》2012年第18期88-95,共8页Mathematics in Practice and Theory

基　　金：教育部人文社科基金(12YJC630248)

摘　　要：提出了一种在对预报因子集进行模糊聚类分析基础上构建径流预测模型的新方法:先通过模糊C-均值聚类将历史径流数据进行分类,然后利用小波神经网络分别建立预报因子集类别变量特征值与观测值之间的局部预测模型,并设计了特征值分类识别器,自动搜寻相适应的局部网络模型进行预测.通过西南某水库2011年日平均入库来流的计算实例对简单小波神经网络预测模型和所建的基于FCM与小波神经网络的预测模型进行了比较,结果较为满意.A new method of a runoff forecast based on fuzzy clustering analysis for fore- casting factor sets is presented. The historical runoff data are divided into four categories using the fuzzy C-means clustering. Then the part forecasting model between forecasting factor＇s class variable characteristic value and observation value is established by use of wavelet neural network model. Designed the categorized recognizer of characteristic value can automatically search adapting network model to forecast. As an example, a comparison is made between single wavelet neural model and forecasting model proposed in this paper. The results show that the forecasting precision of the latter is higher than that of former.

关 键 词：径流预测 组合小波神经网络模型 模糊C均值聚类 遗传算法 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]