Daily and weekly rainfall modelling using Artificial Neural Network
Loading...
Date
2018-08
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
G.B. Pant University of Agriculture and Technology, Pantnagar - 263145 (Uttarakhand)
Abstract
Artificial Neural Network technique has been employed to predict daily and weekly rainfall for Nainital station of Uttarakhand, India. The meteorological data from year (2003 to 2013) were used for training of daily and weekly rainfall prediction models and meteorological data from (2014 to 2018) were used for testing of daily and weekly rainfall prediction models. Gamma test was used for the selection of appropriate input variables of daily and weekly rainfall prediction models. The ANN models were trained using multilayer perceptron with one learning rule i.e. Levenberg - Marquardt and two transfer functions viz. Tanh Axon and Sigmoid Axon. The performance of the models was evaluated qualitatively by visual observation and quantitatively using different performance indices viz. Root Mean Square Error, Correlation Coefficient, Coefficient of Efficiency, Percent Bias and Integral Square Error. It was observed that both Tahn Axon and Sigmoid Axon activation functions are capable of predicting the daily and weekly rainfall with almost equal prediction efficiency, for daily rainfall model structure of 5 – 8 – 8 - 1 with L evenberg – M arquardt and Tahn Axon, and model structure 5 -10 - 10-1 with Levenberg –M arquardt and Sigmoid Axon performed well. And, in case of weekly rainfall the model structure of 5 -10-10-1 was found to be working well in both the cases for Nainital station of Uttarakhand, India.
Description
Keywords
null