AUTHOR: AFUECHETA, EMMANUEL OSITA
AFFILIATION: NNAMDI AZIKIWE UNIVERSITY, AWKA
In this study, we considered the problem of choosing between two statistical techniques, namely logistic regression and linear discriminant analysis which are mostly used for the analysis of categorical outcome variables. While both are appropriate for the development of linear classification models, linear discriminant analysis makes more assumptions about the underlying data. Hence, it is assumed that logistic regression is a more flexible and more robust method in case of violation of these assumptions. In this work, the comparison of the two methods is based on the normality assumption, different sample sizes, and number of variables. The performance of the two techniques is assessed using several measures of predictive accuracy, which are: B-index, Q-index, Cindex, and Classification error (CE). The results obtained show that logistic regression model agrees with linear discriminant analysis to an extent, but logistic regression slightly outperformed linear discriminant analysis especially when data are not normally distributed.
TO VIEW THE FULL CONTENT OF THIS DOCUMENT, PLEASE VISIT THE UNIZIK LIBRARY WEBSITE USING THIS LINK, http://www.naulibrary.org/dglibrary/admin/book_directory/Thesis/10377.pdf
Tags: Asthma Prevalence., Classification Error, Coefficient Estimation, Data Simulation Procedure, Discriminant Analysis, Goodness-of-fit Test, Hosmer=Lemeshow, Logistic Model, Logistic Regression, Logistic Regression Analysis, Parameter Calculation, Statistics Thesis 2010