Rainfall forecasting is an important research problem due to the importance of water for the agricultural, domestic, and industrial sectors, especially when considering extreme weather events caused by climate change. This paper compares the forecasting accuracy of the autoregressive integrated moving average (ARIMA), neural networks (NN) and higher-order Markov models (HOMM) using the observed annual rainfall data series of Tamil Nadu, India. The estimation of all the parameters of the studied model is done by the maximum likelihood method. The prediction accuracy of the models is evaluated using the mean squared error (MSE) criterion. The empirical results show that the NN model provides the highest accuracy in rainfall forecast, followed by ARIMA and higher-order Markov chain models, in that order

Projects

Currently no objects available

Currently no objects available