Performance of the New Ridge Regression Parameters
Mowafaq Muhammed Al-Kassab *
Department of Mathematics Education, Faculty of Education, Tishk International University, Iraq.
Mohammed Qasim Al-Awjar
Department of Statistics and Informatics, College of Computers and Mathematics, Mosul University, Iraq.
*Author to whom correspondence should be addressed.
Abstract
A new approach is presented to find the ridge parameter k when the multiple regression model suffers from multicollinearity. This approach studied two cases, for the value k, scalar, and matrix. A comparison between this proposed ridge parameter and other well-known ridge parameters evaluated elsewhere, in terms of the mean squares error criterion, is given. Examples from several research papers are conducted to illustrate the optimality of this proposed ridge parameter k.
Keywords: Least squares, multicollinearity, ridge parameters, scalar, vector, matrix, mean squared error.