Elastic-Net Regularization in Statistical Downscaling to Estimate Rainfall

Abstract

The impact of climate change and the importance of rainfall as the climate elements so that estimation
of rainfall is important to be implement. The estimation of rainfall used statistical downscaling (SD) technique
which utilize functional relationship approach between local scale data (rainfall) and global scale data (Global
Circulation Model output-GCM). In general, GCM output data are multicollinear, so it is required a technique to
overcome multicollinearity such as elastic-net regularization. The data used are GCM output above Indramayu
regency as explanatory variables and the monthly rainfall data ZOM 79 as response variables. The result shows
that the rainfall estimations with elastic-net at ZOM 79 from 2010 to 2013 were consistent.
Keywords: Elastic-net, multicolinearity, statistical downscaling.