Μια Monte Carlo μελέτη των εκτιμητών Ridge και ελαχίστων τετραγώνων
Part of : Σπουδαί : journal of economics and business ; Vol.41, No.2, 1991, pages 191-204
Issue:
Pages:
191-204
Author:
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.
Subject (LC):
Keywords:
econometrics
Notes:
Περιέχει πίνακες και βιβλιογραφία
