Inflated log-Lindley distribution for modeling continuous data bounded in unit interval with possible mass at boundaries.
In: Chilean Journal of Statistics (ChJS), Jg. 14 (2023-12-01), Heft 2, S. 123-141
Online
academicJournal
Zugriff:
The log-Lindley distribution was introduced as an alternative to the famed beta distribution to model data in the unit interval. The present article introduces inflated versions of the log-Lindley distribution. Distributional properties are investigated. Special cases of the proposed distributions are shown to be members of the exponential family. Moment and maximum likelihood estimation of the parameters are discussed with derivation of the exact expression for the information matrix. Then, the proposed distributions are compared with the inflated unit Lindley and inflated beta distributions for modeling four sets of data regarding school wise pass proportions of high school leaving examinations having zero and/or one inflation. Findings clearly suggest the superiority of the proposed distribution based on the Akaike information criteria and observed and fitted distribution function plots for all the cases considered. [ABSTRACT FROM AUTHOR]
Titel: |
Inflated log-Lindley distribution for modeling continuous data bounded in unit interval with possible mass at boundaries.
|
---|---|
Autor/in / Beteiligte Person: | BHATTACHARJEE, SAHANA ; CHAKRABORTY, SUBRATA |
Link: | |
Zeitschrift: | Chilean Journal of Statistics (ChJS), Jg. 14 (2023-12-01), Heft 2, S. 123-141 |
Veröffentlichung: | 2023 |
Medientyp: | academicJournal |
ISSN: | 0718-7912 (print) |
DOI: | 10.32372/ChJS.14-02-04 |
Schlagwort: |
|
Sonstiges: |
|