RAINFALL FORECASTING MODEL USING ADALINE AND REGRESSION ALGORITHM

Abstract

High rainfall intensity could cause flooding or inundation which needs to be forecasted to estimate the amount of rainfall that will come. Forecasting is the process of making future predictions based on past and present data and most commonly with trend analysis. Computational models can be used using artificial statistics and statistics. The advantage of forecasting using artificial intelligence is able to calculate with non-linear data and limited parameters while statistical forecasting is the opposite. Adaptive Linear Neuron (ADALINE) is one of the learning algorithms of the artificial neural network algorithm. The learning algorithm is able to recognize historical data patterns so that it can provide future pattern predictions. This study have compared the performance of computational models using ADALINE artificial intelligence with statistical computing models, namely regression. The rainfall data used was the rainfall data in South Kalimantan Province 2008-2013. Comparison of the performance fro the two algorithms was using root mean square error (RMSE). The smaller the RMSE value was more better. The results of the two algorithms has been obtained RMSE ADALINE algorithm was 0.0729 while RMSE REGRESSION algorithm was 0.0900. This shown that the performance of the ADALINE algorithm was better than the REGRESSION algorithm for forecasting rainfall data in the Province of South Kalimantan in 2008-2013.