Open AccessPENERAPAN REGRESI QUASI-LIKELIHOOD PADA DATA CACAH (COUNT DATA) YANG MENGALAMI OVERDISPERSI DALAM REGRESI POISSON
Author(s) -
Desak Putu Prami Meitriani,
I Komang Gde Sukarsa,
I Putu Eka Nila Kencana
Publication year - 2013
Publication title -
e-jurnal matematika
Language(s) - English
Resource type - Journals
ISSN - 2303-1751
DOI - 10.24843/mtk.2013.v02.i02.p036
Subject(s) - overdispersion , quasi likelihood , poisson regression , count data , statistics , poisson distribution , regression analysis , mathematics , zero inflated model , regression , generalized linear model , standard error , econometrics , medicine , population , environmental health
Poisson regression can be used to analyze count data, with assuming equidispersion. However, in the case of overdispersion often occur in the count data. The implementation of Poisson Regression can not be applied on this data because the data having overdispersion, that will lead to underestimate the standard error. Thus, use Quasi-Likelihood regression on this data. Quasi-Likelihood regression was also could not handle the overdispersion, but Quasi-Likelihood regression can improve the value of the standard error becomes greater than the value of the standard error on Poisson regression. Thus, by using the Quasi-Likelihood regression obtained three independent variables that affect the number of divorce cases in each urban city of Denpasar in 2011.