Application of the local influence analysis in
logistic regression models

Édila Cristina de SOUZA[1]

Edwin Moises Marcos ORTEGA1

Vicente Garibay CANCHO[2]

§     ABSTRACT: An important stage after the formulation and adjustment of a regression model is the diagnosis analysis. Logistic regression is one of the main methods for modeling data and even when the response of interest is not originally of the binary type, some researchers have dichotomized the response in a way that the success probability can be modeled through logistic regression. The estimators obtained for the maximum likelihood and Bayesian methods were used. In this study we consider a study of diagnostic methods with logistic regression, using the local influence technique of Cook (1986). We investigate the application of the local influence technique under different types of disturbance. As an illustration, we show the application of the developed results obtained with real data sets.

§     KEYWORDS: Logistic regression;  diagnostic analysis;  local influence; Bayesian method.



[1]Departamento de Ciências Exatas, Universidade de São Paulo - ESALQ/USP, CEP 13418-900 Piracicaba, SP, Brasil. E-mail: edwin@esalq.usp.br

[2] Instituo de Ciências Matemáticas e de Computação, Universidade de São Paulo – USP, Caixa Postal 668, CEP: 13560-970, São Carlos, SP, Brasil. E-mail: gariby@icmc.usp.br