Time-series modelling and forecasting: modelling of rainfall prediction using ARIMA model

This work presents a Time Series Modeler (TSM) for forecasting the rainfall of a coastal region in India. In developing this model, a five-year dataset (2009-2013) consisting of temperature, dew point, wind speed, max. temperature, min. temperature, visibility and rainfall are considered as prime attributes. As a novel attempt, TSM of Statistical Package for Social Studies (SPSS) has been applied for training and testing this dataset. The performance criteria for the evaluation of this model are evaluated based on the significant values of the statistical performance measures namely mean absolute deviation (MAD), mean squared error (MSE), mean absolute percent error (MAPE) and root mean squared error (RMSE) and therefore a reliable model for rainfall prediction is possible. The results obtained through this model are well acceptable with the prediction accuracy range of 80%. This model is built on auto regressive integrated moving average (ARIMA) model of TSM in SPSS 20.0.