ANALYSIS OF MERIDIAN ESTIMATOR PERFORMANCE FOR NON-GAUSSIAN PDF DATA SAMPLES
Dmitriy Kurkin, Alexey A. Roenko, Владимир Васильевич Лукин, Igor Djurović · Telecommunications and Radio Engineering · 2010
A sample meridian estimator of location parameter (LP) has been proposed recently and shown to be robust and controllable by means of a tunable parameter δ. The estimator properties have been initially studied but not analyzed thoroughly. In this paper we address several practical questions. First, we analyze conditions under which statistical properties of meridian estimator of LP considerably differ from those ones of the sample median estimator. Second, we give examples of probability density functions for which the sample meridian estimates can be sufficiently more accurate than sample median. Third, we consider practical situations when useful properties of the meridian estimator of LP can be exploited in practice of signal and data processing.