Journal Article

The Performance of the Modified Detection Method in the Presence of High Leverage Collinearity Enhancing Observations

Abstract

The presence of outliers and multicollinearity are inevitable in real data sets and they have an unduly effect on the parameter estimation of multiple linear regression models. It is now evident that outliers in the X-direction or high leverage points are another source of multicollinearity. These leverage points may enhance multicollinearity in a dataset. We call this high leverage collinearity enhancing observations. However, the detection method LTSR-HLCIM which is based on DRGP(MVE) depends on minimum volume ellipsoid which has a long computing running time, computational complexity and the presence of swamping and masking effects. Another diagnostic measure was introduced to reduce the three short comings mentioned in the existing method by using DRGP(RFCH) instead of DRGP(MVE) used by LTSR-HLCIM. We study several criteria such as sample sizes, percentages and position of high leverage points which cause these leverages to change the multicollinearity pattern of collinear data sets. The results of simulation 0073tudy and real dataset indicates that the modified detection method out-performed the existing method in term of having the highest percentage of correct detection of good and bad leverage collinearity enhancing observation.

Download pdf