An investigative resolution of collinearity through location shift and partial regression splines with an application to crime-data modeling

P. Sangli Therasanathan¹ , Martin L. William²

Section:Research Paper, Product Type: Isroset-Journal
Vol.6 , Issue.1 , pp.53-60, Feb-2019

CrossRef-DOI:   https://doi.org/10.26438/ijsrmss/v6i1.5360

Online published on Feb 28, 2019

Copyright © P. Sangli Therasanathan, Martin L. William . This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
 

View this paper at   Google Scholar | DPI Digital Library

XML View     PDF Download

How to Cite this Paper

416 Views    408 Downloads    111 Downloads

Abstract :
Collinearity is an issue that plagues any regression model primarily built to explain the effect of regressors on a response variable. Various solutions have been attempted in the literature on addressing the collinearity issue, like taking non-linear transformations and dropping a regressor that has a collinear relation with another. But, there are situations where a regressor is affected by collinearity but not exactly in relation to another. Dropping of such a variable leads to complete loss of information of an ‘important` factor and so a different solution is called for in such situations. This paper explores a heuristic approach to resolve collinearity by making a location shift on the affected variable(s) and creating a regression spline. This approach is illustrated for modeling district-wise data on crime against women recorded in 2011 across the districts of the South Indian states, using regressors available from the 2011census data and from data sources on the presence of schools and police stations.

Key-Words / Index Term :
Collinearity, Location shift, Regression splines

References :
[1] D.C. Montgomery, E.A. Peck and G.G. Vining. “Introduction to Linear Regression Analysis”, John Wiley & Sons, New Jersey, USA, pp. 325-381, 2012
[2] D.E. Farrar and R.R. Glauber. “Multicollinearity in regression analysis: the problem revisited”, Review of Economics and Statistics, Vol. 49, pp. 92-107, 1967
[3] R. Mason and J.T. Webster. “Regression Analysis and Problems of Multicollinearity”, Communications in Statistics, Vol. 4, Issue 3, pp. 277-292, 1975
[4] A.C. Harvey. “Some Comments on Multicollinearity in Regression”, Journal of the Royal Statistical Society, Series C (Applied Statistics), Vol. 26, Issue 2, pp. 188-191, 1977
[5] G.W. Stewart. “Collinearity and Least Square Regression”, Statistical Science. Vol. 2, Issue 1, pp. 68-94, 1987
[6] J. Fox and G. Monette. “Generalized Collinearity Diagnostics”, Journal of the American Statistical Association, Vol. 87, pp. 178-183, 1992
[7] C.F. Mela and P.K. Kopalle. “The impact of collinearity on regression analysis: the asymmetric effect of negative and positive correlations”, Applied Economics, Vol. 34, pp. 667-677, 2002
[8] M. Shacham and N. Brauner. “Minimizing the effects of collinearity in polynomial regression”, Ind. Eng. Chem. Res. Vol. 36, pp. 4405–4412, 1997
[9] J.H. Friedman. “Multivariate Adaptive Regression Splines”, The Annals of Statistics, Vol. 19, Issue 1, pp.1-67, 1991