CLASSIFICATION BINARY MODELS FOR BIOMEDICAL DATA: SIMPLE PROBABILISTIC NETWORKS AND LOGISTIC REGRESSION

Main Article Content

Anderson ARA
Francisco LOUZADA
Luis Aparecido MILAN

Abstract

In the biomedical area a critical factor is whether a classication model is accurate enough in order to provide correct classication whether or not a patient has a certain disease. Several techniques may be used in order to accommodate such situation. In this context, Bayesian networks have emerged as a practical classication technology with successful applications in many elds. At the same time, logistic regression is a widely used statistical classication method and evidenced in the literature. In the current paper we focus on investigating the preditive performance of a probabilistic networks in its simple particular case, the so called naive Bayes network, compared to the logistic regression. A systematic simulation study is performed and the procedures are illustrated in some benchmark biomedical data sets. data sets.

Article Details

How to Cite
ARA, A., LOUZADA, F., & MILAN, L. A. (2018). CLASSIFICATION BINARY MODELS FOR BIOMEDICAL DATA: SIMPLE PROBABILISTIC NETWORKS AND LOGISTIC REGRESSION. Brazilian Journal of Biometrics, 36(1), 48–55. https://doi.org/10.28951/rbb.v36i1.114
Section
Articles