Master'sOpen Access

The comparison of log-linear modelling and logistic regression

2012
0 views
0 downloads
Advisor: Prof. Dr. Semra Erbaş

Abstract (EN)

With the growing interest in social researchs, categorical data analysis has become more important. In this study logistic regression, lojit models and log-linear models of categorical data analysis methods are discussed and compared with the statistical properties differently. Researchs? selection criteria for lojistic regression or log-linear models were examined and the two applications were applied on the Turkish Statistical Institute 2009 Household Budget Survey data. In the first application, variables of the model in the final stage were compared by log-linear models and logistic regression. In the second application, the relationship between the log-linear models and logit models were discussed and for two and three-way contingency tables logit model parameter estimates are shown to be equal to the log-linear model parameter estimates.Key Words :Categorical Data, contingency tables, logistic regression, lojit model, log-linear models

Author

Dr. Gülsüm Merve Oktay

How to Cite

Gülsüm Merve Oktay (Master Thesis). The comparison of log-linear modelling and logistic regression, 2012, Gazi University.

License

Tüm Hakları Saklıdır

This work is shared under the specified license terms.

More theses from Gazi University