کد مقاله کد نشریه سال انتشار مقاله انگلیسی نسخه تمام متن
6412824 1629933 2014 12 صفحه PDF دانلود رایگان
عنوان انگلیسی مقاله ISI
Comparative study of different wavelet based neural network models for rainfall-runoff modeling
موضوعات مرتبط
مهندسی و علوم پایه علوم زمین و سیارات فرآیندهای سطح زمین
پیش نمایش صفحه اول مقاله
Comparative study of different wavelet based neural network models for rainfall-runoff modeling
چکیده انگلیسی


- 23 Mother wavelet functions are used to transform input rainfall data.
- 92 ANN models with transformed input rainfall data are developed.
- db2 Wavelet function performed best with Continuous Wavelet Transformation (CWT).
- db8 Wavelet function performed best with Discrete Wavelet Transformation (DWT).
- The DWT gives best result at maximum possible level of decomposition of the input data.

SummaryThe use of wavelet transformation in rainfall-runoff modeling has become popular because of its ability to simultaneously deal with both the spectral and the temporal information contained within time series data. The selection of an appropriate wavelet function plays a crucial role for successful implementation of the wavelet based rainfall-runoff artificial neural network models as it can lead to further enhancement in the model performance. The present study is therefore conducted to evaluate the effects of 23 mother wavelet functions on the performance of the hybrid wavelet based artificial neural network rainfall-runoff models. The hybrid Multilayer Perceptron Neural Network (MLPNN) and the Radial Basis Function Neural Network (RBFNN) models are developed in this study using both the continuous wavelet and the discrete wavelet transformation types. The performances of the 92 developed wavelet based neural network models with all the 23 mother wavelet functions are compared with the neural network models developed without wavelet transformations. It is found that among all the models tested, the discrete wavelet transform multilayer perceptron neural network (DWTMLPNN) and the discrete wavelet transform radial basis function (DWTRBFNN) models at decomposition level nine with the db8 wavelet function has the best performance. The result also shows that the pre-processing of input rainfall data by the wavelet transformation can significantly increases performance of the MLPNN and the RBFNN rainfall-runoff models.

ناشر
Database: Elsevier - ScienceDirect (ساینس دایرکت)
Journal: Journal of Hydrology - Volume 515, 16 July 2014, Pages 47-58
نویسندگان
, , ,