Multiple-Group Logistic Regression Diagnostics
Journal of the Royal Statistical Society Series C (Applied Statistics)1989Vol. 38(3), pp. 425–425
Citations Over TimeTop 11% of 1989 papers
Abstract
SUMMARY The regression diagnostics introduced by Pregibon for the dichotomous logistic model are extended to multiple groups viewed as a multivariate generalized linear model. We develop diagnostics which measure the influence of each observation on the performance of the estimated classification rule, thus highlighting possible multivariate outlying and/or influential observations. As an illustration, the diagnostics are applied to enzyme data from liver disease patients. It is shown that multivariate regression diagnostics constitute an indispensable tool for the practitioner to construct appropriate polychotomous logistic models.
Related Papers
- → A Generalized Hosmer–Lemeshow Goodness-of-Fit Test for Multinomial Logistic Regression Models(2012)270 cited
- Logistic regression model and its application(2012)
- APPLICATION OF LOGISTIC REGRESSION MODELS IN RISK MANAGEMENT(2021)
- → Binary Logistic Regression(2016)
- → Collapsibility and Suppression for Cumulative Logistic Model(2005)