Resumo

MAHDIZADEH, MAHDI. On the Use of Ranked Set Samples in Entropy Based Test of Fit for the Laplace Distribution. Rev.Colomb.Estad. [online]. 2012, vol.35, n.3, pp.443-455. ISSN 0120-1751.

Statistical methods based on ranked set sampling (RSS) often lead to marked improvement over analogous methods based on simple random sampling (SRS). Entropy has been influential in the development of measures of fit of parametric models to the data. This article develops goodness-of-fit tests of the Laplace distribution based on sample entropy when data are collected according to some RSS-based schemes. For each design, critical values of the corresponding test statistic are estimated, by means of simulation, for some sample sizes. A Monte Carlo study on the power of the new tests is performed for several alternative distributions and sample sizes in order to compare our proposal with available method in SRS. Simulation results show that RSS and its variations lead to tests giving higher power than the test based on SRS.

Palavras-chave : Entropy estimation; Goodness-of-fit test; Ranked set sampling.

        · resumo em Espanhol     · texto em Inglês     · Inglês ( pdf )