المستخلص: |
Several biased estimator have bean proposed as alternative to the least squared estimator when multicollinearity is presenting the multiple linear regression model. In this article a class of biased linear estimator of the vector unknown regression coefficient, when multicollinearity among the columns of the design matrix exists, is obtained. The ordinary ridge regression, principal components, partitioned regression and shrinkage estimators are members of this class. The theoretical properties of this new estimator were derived.
|