Penerapan Jaringan Saraf Tiruan untuk Pemodelan Prakiraan Curah Hujan Bulanan

Abstract

Rain is very usefull for life but heavy rainfall will lead to disaster if not anticipated properly. Rainfall forecasting is very usefull to anticipate the occurrence of extreme weather. This study developed artificial neural network models for the next three months rainfall forecasting. The data used in this study was the climate data from 1983 to 2013. The climate data used as model input were rainfall, temperature and relative humadity (RH). Temperature and RH averaged on a month period and observed three times a day. The best result from this study was the model with the highest correlation value. The Correlation values of the best model for rainfall forecasting of the next three months were 0.78, 0.83, 0.77 respectively.