New Test Statistics To Assess The Goodness-Of-Fit Of Logistic Regression Models
Loading...
Date
2013-04
Authors
Hussain, Jassim Nassir
Journal Title
Journal ISSN
Volume Title
Publisher
Universiti Sains Malaysia
Abstract
The binary (or binomial) logistic regression model (LRM) is one of the
generalised linear models (GLMs). It is used when the dependent variable is
dichotomous and the independent variables are of any type. LRM are popular in
many applications and in different disciplines including biomedical and social
sciences. Assessing the goodness-of-fit (GOF) is considered to be the important step
after fitting the model to show the adequacy of the LRM in fitting the observations.
The GOF test is defined as an evaluation of how well the estimated outcomes agree
with the observed data. Two techniques may be used to construct the GOF test
statistics of chi-square type. The first technique is based on ungrouped observations.
This technique is not preferred in the LRM for many reasons including that the
obtained distribution and values are incorrect. The second technique is the
preferred technique where it is based on grouping the observations.
Description
Keywords
Assess The Goodness-Of-Fit , Logistic Regression Models