Modeling of Rainfall Prediction over Myanmar Using Polynomial Regression

Abstract

Myanmar is an agricultural country and its economy is largely based upon crop productivity. The occurrence of extreme precipitation variability may lead to significantly reduce crop yields and extensive crop losses. Thus, rainfall prediction becomes an important issue in Myanmar. Regression has since long been a major data analytic tool in many scientific such as behavioral sciences, social sciences, biological sciences, medical sciences, psychometrics and econometrics for predicting. Multi variables polynomial regression (MPR) is one of the statistical regression method used to describe complex nonlinear input output relationships. In this paper, MPR is applied to implement the precipitation forecast model over Myanmar. Myanmar receives its annual rainfall during the summer monsoon season which starts in June and end in September. The model output result is station wide monthly and annual rainfall amount during summer monsoon season. The proposed model results are compared with the result produced by multiple linear regression model (MLR). From the experimental results, it is observed that using MPR method achieves closer agreement between actual and estimated rainfall than using MLR.