Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/288287 
Year of Publication: 
2020
Citation: 
[Journal:] Metrika [ISSN:] 1435-926X [Volume:] 83 [Issue:] 7 [Publisher:] Springer [Place:] Berlin, Heidelberg [Year:] 2020 [Pages:] 853-868
Publisher: 
Springer, Berlin, Heidelberg
Abstract: 
A spacings-based prediction method for future upper record values is proposed as an alternative to maximum likelihood prediction. For an underlying family of distributions with continuous cumulative distribution functions, the general form of the predictor as a function of the estimator of the distributional parameters is established. A connection between this method and the maximum observed likelihood prediction procedure is shown. The maximum product of spacings predictor turns out to be useful to predict the next record value in contrast to likelihood-based procedures, which provide trivial predictors in this particular case. Moreover, examples are given for the exponential and the Pareto distributions, and a real data set is analyzed.
Subjects: 
Point prediction
Cumulative hazard rate
Spacings
Maximum observed likelihood predictor
Upper record values
Exponential distribution
Pareto distribution
JEL: 
F99
M20
Persistent Identifier of the first edition: 
Creative Commons License: 
cc-by Logo
Document Type: 
Article
Document Version: 
Published Version

Files in This Item:
File
Size





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.