Comparison of Machine Learning Models For Rainfall Forecasting
Nazli Bin Mohd Khairudin, Norwati Mustapha, Teh Noranis Binti Mohd. Aris, Maslina Zolkepli
Abstract
Extreme rainfall can lead to a flood occurrence that give a devastating impact on human lives including the agriculture sectors. Accurate rainfall forecasting is crucial in minimizing the consequences derived from the flood. In this study, the rainfall forecast is estimated using 5 different machine learning models which are Artificial Neural Network (ANN), Support Vector Regression (SVR), Decision Tree (DT), Random Forest Algorithm (RFA), and Long Short-Term Memory (LSTM). Average weekly rainfall data of Kuala Krai rainfall station have been used as predictor variable for this study. The performances of the modelling approaches are evaluated by the statistical score metrics of root mean squared error (RMSE) and mean absolute error (MAE). The results have shown that LSTM performed best among other models in forecasting the rainfall of Kuala Krai station.