![]() | يوجد فقط 14 صفحة متاحة للعرض العام |
المستخلص One of the most important statistical properties of least squares is the property of unbiased least squares estimator for the parameter model as if it was biased estimate for the parameter, this will lead to the erroneous prediction of the dependent variable and leads to a false statistical inferences or misleading. This paper addresses the problem of estimation of predictive regressions, by which the time series of the dependent variable is regressed on the lagged time series of the independent variable. The independent variable is an AR (1) process and its residuals are correlated with the predictive regression’s residuals, the ordinary least squares (OLS) estimator of the independent variable’s coefficient, will be biased in finite sample. The aim of this study has access to the best method to reduce the bias of the methods available to reduce the bias by evaluating these methods by using some of statistical criteria and application on the data of Egyptian Company for Mobile Services (Mobinil). The results of this study are that the Lewellen’s Method at value ( ) was the best method to reduce the bias. We have confirmed that result by using the simulation for the data and compare the methods used to reduce the bias of these data by using the standards previously and the result was similar as a result of actual data |