REVIEW AND CLASSIFICATIONS OF THE RIDGE PARAMETER ESTIMATION TECHNIQUES

Ridge parameter estimation techniques under the inuence of multi-collinearity in Linear regression model were reviewed and classified into different forms and various types. The different forms are Fixed Maximum (FM), Varying Maximum (VM), Arithmetic Mean (AM), Geometric Mean (GM), Harmonic Mean (HM) and Median (M) and the various types are Original (O), Reciprocal (R), Square Root (SR) and Reciprocal of Square Root (RSR). These classications resulted into proposing some other techniques of Ridge parameter estimation. Investigation of the existing and proposed ones were done by conducting 1000 Monte-Carlo experiments under five (5) levels of multicollinearity ( $\rho=0.8, 0.9, 0.95, 0.99, 0.999$), three (3) levels of error variance ($\sigma^2=0.25,1,25$) and five levels of sample size ($n=10,20,30,40,50$). The relative efficiency ($RF\leq 0.75$) of the techniques resulting from the ratio of their mean square error and that of the ordinary least square was used to compare the techniques.  Results show that the proposed techniques perform better than the existing ones in some situations; and that the best technique is generally the ridge parameter in the form of Harmonic Mean, Fixed Maximum and Varying Maximum in their Original and Square Root types.