Forecasting Rainfall Distribution Based on Deseasonalising Fuzzy Time Series Model

Abstract

Rainfall prediction is an essential process to reduce loss of lives and properties. However, the accuracy of this prediction has been of many concerns in literature. Therefore, this paper proposed a model of rainfall prediction based on deseasonalising data and fuzzy time series concept. The historical data of rainfall distribution were collected from Drainage and Irrigation Department, Perlis Malaysia between January 2000 and December 2013. These data were analysed in order to determine the seasonal components using fuzzy time series as a medium. The study made use of deseasonalising rainfall data by employing fuzzy time series model in order to forecast the rainfall distribution. The model performance was evaluated by using statistical criteria of Mean Squared Error (MSE) and Root Mean Squared Error (RMSE). The obtained result was compared with several forecasting models in literature and it was found to be more accurate than others. Hence, this study demonstrates that fuzzy time series model is more suitable for the accurate prediction of rainfall distributions.