Hail Prediction Model Using Binary Logistic Regression In West Java Region Bella Suci Niati
Department of Physics, Faculty of Mathematics and natural Sciences, Sriwijaya university, Inderalaya, Indonesia
Abstract
The hail phenomenon in Indonesia has increased in recent years, especially in the western part of Java. Recorded from 2010 to 2018 as many as 27 cases of hail events. However,hail has not been well predicted until now so that a system for predicting hail events needs to be made. This study aims to create a model using binary logistic regression for find the relationship between factors that affect the occurrence of hail. The data used includes radar data and top air observation from 2010-2018. The location of this research in DKI Jakarta, Banten and West Java. To determine the most suitable binary logistic regression model,logistic regression analysis is done by test each parameter coefficient and describe descriptive. Based on the significant parameter test by using partial test, variable of ZMax and 500mb zonal wind which had significant effect to hail.Logistic regression model that had the smallest statistical value of AIC was the most feasible model to use. Results showed that the best model was obtained with the statistic AIC test of 41.88971.And then the logistic regression was used to determine risk/no risk of hail. The results show that the forecast models each have POD,CSI and FAR of 1,0,and 1.
