Comparisons of the LH Moments and the L Moments

Abstract

This paper discusses comparisons of the LH moments method with L moments method. LH moments, a generalization of L moments, based on linear combinations of higher-order statistics was introduced for charactering the upper part of distributions and larger events in data by Wang (1997). Analysis of observed data shows that using LH moments estimates of the upper part of distribution events are expected to be more reasonable than the L moments estimates. A comparison of the LH moment diagram and the L moment diagram of the data also shows that the GEV distribution describes the LH moment ratios better than the L moments.