Journal Article

Application Of Simulation Technique to Manage Collinearity

Abstract

The presence of multicollinearity is unavoidable in a data set. It is obvious that high leverage points are another source of multicollinearity. These leverage points may reduce multicollinearity in a dataset. We call this high leverage collinearity reducing observations. Several criteria were studied to assess the merit of the proposed detection method over the existing method using different sample sizes, percentages and position of high leverage points which cause these leverages to change the multicollinearity pattern of non-collinear data set. 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 shortcomings mentioned in the existing method by using DRGP(RFCH) in (Improvised LTSR-HLCIM) instead of DRGP(MVE) used by LTSRHLCIM. The results of simulation study indicated that the modified detection method outperformed the existing method in term of having the highest percentage of correct detection of good and bad leverage collinearity enhancing observations of the small, medium and large sample sizes with three different levels of HLPs.

Download pdf