COMPARISON OF ARIMA AND ANN FOR FORECASTING THE ANNUAL RAINFALL OF NADIA DISTRICT, WEST BENGAL, INDIAK. Sathees Kumar, T. Gowthaman and Banjul BhattacharyyaAbstract Every fluctuation in rainfall and temperature will affect crop yields since Indian agriculture is very susceptible to climate fluctuations, especially to temperature and rainfall. Planning and management of natural resources requires an understanding of the geographical and temporal distribution and changing trends in climatic factors. In order to better understand the variability pattern in climate data and perhaps even forecast short- and long-term changes in the series, time series analysis can be a very useful technique. The annual rainfall records for the Nadia district of west bengal from 1981 to 2021 have been examined in this study. The rainfall data were modelled using linear parametric technique Autoregressive Integrated Moving Average (ARIMA) and nonlinear nonparametric technique Artificial Neural Network (ANN). Model performance of ARIMA and ANN were compared. Result revealed that ARIMA was performed better than ANN for forecasting the rainfall of Nadia district. This forecasts from ARIMA are anticipated to assist decision-makers in the effective scheduling of agricultural management, urban planning, rainfall collection, and flood prediction. |