Μια Monte Carlo μελέτη των εκτιμητών ridge και ελαχίστων τετραγώνων

Abstract

Multicollinearity is a very severe problem in many statistical and econometrics application.
The most promising technique for reducing the harmful effect of multicillinearity is called Ridge
Regression (RR).
At this study ("A MONTE CARLO STUDY OF RIDGE AND OLS ESTIMATORS") a Monte-Carlo
experiment was used for the OLS estimator and Ridge estimators which were determined by five different
methods of estimations of "k". These estimators were compered using the mean squared error (MSE) and
the effectiveness index (EFF) criteria. The Monte-Carlo experiment was based on two models from the
Greek economy.