A COMPARISON OF SELECTION RULES OF k IN RIDGE ESTIMATORS

Abstract

Least squares estimates of the parameters in the usual linear regression model have a high probability of being unsatisfactory when the predictor variables are multicollinear. HOEAL and KENNARD have demonstrated that these undesirable effect of multicollinearity can be reduced by using ridge estimates in placed of the least squares estimates. Ridge regression theory in literature has been mainly concerned with selection of k. In this thesis, we demonstrate a simple selection rules of optimal k and compare this rule and the rules in literatures.;이 논문에서는 HOEAL 과 KENNARD 가 제시한 능형회귀추정량 (β(k))의 모수 k를 최적화 하는 방법을 TMSE 기준 하에서 제시하고, 이 방법과 문헌에 나와 있는 방법들을 예를 통해 비교하였다.