Generalized least-squares estimators for the thickness of heavy tails

Academic Article

Abstract

  • For a probability distribution with power law tails, a log-log transformation makes the tails of the empirical distribution function resemble a straight line, leading to a least-squares estimate of the tail thickness. Taking into account the mean and covariance structure of the extreme order statistics leads to improved tail estimators, and a surprising connection with Hill's estimator. © 2002 Elsevier B.V. All rights reserved.
  • Digital Object Identifier (doi)

    Pubmed Id

  • 25734998
  • Author List

  • Aban IB; Meerschaert MM
  • Start Page

  • 341
  • End Page

  • 352
  • Volume

  • 119
  • Issue

  • 2