Unbiased L1 and L∞ estimation |
| |
Authors: | RW Farebrother |
| |
Institution: | Department of Econometrics and Social statistics , University of Manchester , Manchester , U.K., M13 9PL |
| |
Abstract: | Sielken and Heartely 1973 have shown that the L1 and L∞ estimation problems may be formulated in such a way as to yield unbiased estimators of in the standard linear model y = Xβ + ε In this paper we will show that the L1 estimation problem is closely related to the dual of the L∞ estimation problem and vice versa. We will use this resu;t to obtain four fistiner lineat programming problems which yield unbiased L1 and L∞ estimators of β. |
| |
Keywords: | Lp estimators linear prograrmming regression quantiles symmetric distributions |
|
|