Rainfall Prediction Due to El Nino Factors Using Recurrent Neural Networks

Abstract

El Nino is one of the natural phenomena that have a significant influence on the weather, causing a longer dry season in several regions of Indonesia, one of which is the city of Lampung. One way to anticipate a long drought is to predict rainfall, by looking at the intensity of the rain. This paper proposes rainfall prediction using a recurrent neural network. Weather variables used to predict rainfall include air humidity, wind speed obtained from BMKG stations, and SOI index obtained from the NCDC website in the past 10 years. Weather data will be interpolated and extracted to find the maximum weather value per 4 weeks, the next step is overlapping, after which the data segmentation and normalization become 0-1 to make the data values not far adrift. The results showed the prediction of rainfall with a vulnerable 4 weeks using the Recurrent Neural Networks method produces an accuracy of 89.53%.