Your Cart

Your cart is empty

Home  >  Volume 27 (July 2014)

28. The Application of Quantile Regression Estimator as Alternative to Ordinary Least Squares by A. H. Bello. Volume27, (July, 2014), pp 227 – 234
Sale price: $5.00


This research work attempts to study the robustness of quantile regression as alternative to ordinary least squares method. Whereas the sum of squared errors is minimized in ordinary least squares regression, the median regression estimator minimized the sum of absolute errors, in quantile regression. The remaining conditional quantile functions are estimated by minimizing an asymmetrically weighted sum of absolute errors. A linear multiple regression model was fitted for both ordinary least squares and quantile regression, the results of the parameters estimated from analysis of economic growth rate of Gross Domestic Product (GDP) in Nigeria for thirty one year’s which were computed with the use of Stata (12.0), MS Excel and SPSS were compared and it was found that the existence of outliers and non-normality assumption are violated when Ordinary least squares estimates(OLS) was used and the result can be misleading. Quantile regression gives true relationship between response variables and covariate for different subsections of the sample and therefore gives a better and efficient estimate of the relationship among random variables..

    Keywords: Quantile Regression, Ordinary least squares, Outliers, Gross Domestic Product, Nigeria.