Hippocampus using regression to find the most appropriate model to represent Meteorological data in the city of Arbil during the period (1998-2010)

Abstract

In this paper we shall explain how to use Robust Regression methods and their application to estimate regression parameters in order to fit the best regression model in case of existence outlier observations in the data, The outliers were detected using statistical methods, DIFFT (Difference Fit), Leverage, and Box-Whisker-Plot tests. Then two other statistical models were used to estimate linear regression model, Least Median Squares (LMS) and Least Trimmed Squares (LTS). The results of these two modes then were compared with results obtained using Ordinary Least Squares (OLS) model in order to find the best regression model. We have used meteorological data in this research. Finally the test we used for comparing these three models is MSE statistic.