https://scholars.lib.ntu.edu.tw/handle/123456789/87519
標題: | 結合OLS與SGA建構輻狀基底類神經網路於洪水預測之研究 Building Radial Basic Function Neural Network by Integrating OLS and SGA for Flood Forecasting |
作者: | 張麗秋 林永堂 張斐章 Chang, Li-Chiu Lin, Yung-Tang Chang,Fi-John |
關鍵字: | 降雨-逕流模式;輻狀基底函數神經網路;垂直最小平方法;序率坡降法;倒傳遞類神經網路;Rainfall-runoff model;Radial basis function neural network;Orthogonal least squares;Stochastic gradient approach;Back propagation neural network | 公開日期: | 十二月-2005 | 卷: | 53 | 期: | 4 | 起(迄)頁: | 25-38 | 來源出版物: | 臺灣水利 | 摘要: | 本研究以輻狀基底函數類神經網路來架構集水區降雨-逕流模式。建構輻狀基底函數類神經網路可分為兩階段,前階段以垂直最小平方法(OLS)決定隱藏層中輻狀基底函數之個數及其中心位置;後階段則以序率坡降法(SGA)推求網路隱藏層與輸出層之參數。OLS演算法能有系統地從輸入向量中挑選出影響推估結果最大的輸入資料當作隱藏層的神經元;而SGA則利用最陡坡降法的概念,修正網路之參數。我們首先以合成的傅立葉函數序列來測試OLS搭配SGA之輻狀基底函數類神經網路(RBFNN)的技術性,結果顯示其具有良好的推估能力與準確性;其次,以蘭陽溪降雨-逕流資料為例,並與倒傳遞類神經網路(BPNN)進行比較,結果顯示輻狀基底函數類神經網路比倒傳遞類神經網路之預測效能更佳,且能夠精確推估下一時刻與下二時刻的洪水事件。 In this study, the radial basis function neural network (RBFNN) is used to model the rainfall-runoff process. The training process of RBFNN includes two phases. First, the Orthogonal Least Squares (OLS) is used to determine the number and the center of radial basis function in the hidden layer. Then, the parameters in radial basis functions and the connected weights between the hidden layer and output layer are determined by the Stochastic Gradient Approach (SGA). The OLS algorithm could systematically identify effective input data and set them as the nodes of hidden layer, while the SGA algorithm could search optimal parameters of the network. The proposed RBFNN is first verified by using a theoretical Fourier function. The results show that the model has great ability and high accuracy in simulation of the theoretical case. To further investigate the models' sapplicability, Lanyang River is used as case study. The Back propagation neural network (BPNN) is also performed for the purpose of comparison. The results demonstrate that the proposed RBFNN has much better performance than BPNN. RBFNN not only provides an efficient way to model the rainfall-runoff process, but also give precise one-step and two-step ahead flood forecasts. |
URI: | http://ntur.lib.ntu.edu.tw//handle/246246/193435 |
顯示於: | 生物環境系統工程學系 |
檔案 | 描述 | 大小 | 格式 | |
---|---|---|---|---|
結合 OLS 與 SGA 建構輻狀基底類.pdf | 851.31 kB | Adobe PDF | 檢視/開啟 |
在 IR 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。