Rainfall Forecasting using the Classification and Regression Tree (CART) Algorithm and Adaptive Synthetic Sampling (Study Case: Bandung Regency)

Abstract

Indonesia is a country that can experience potentially adverse climate change. More than 50% of the population in Bandung Regency works in the agricultural sector. Hence, the prediction of rainfall is essential in agriculture to produce the best harvest and to minimize losses. In this study, a Classification and Regression Tree (CART) algorithm were used to forecast the rainfall in Bandung Regency. Furthermore, an Adaptive Synthetic Sampling (ADASYN) algorithm was added to optimize the model produced due to a class imbalance in the data. The weather data was collected from the Meteorology, Climatology and Geophysics Agency of Indonesia (BMKG) from 2005–2017. The results showed that using the CART algorithm yielded 93.94% rainfall prediction accuracy with a 1.38 s running time whereas using ADASYN and CART yielded an accuracy of 98.18% with a 1.48 s running time.