Logistic regression is used for modeling a categorical variable (for example yes/no) in terms of a set of covariates. The intention is to assess the impact of these covariates on the conditional probabilities estimated by the model.
The problem is that this technique is not robust to outliers, and even minor ones can greatly impact our estimation. This is the same problem that we had for linear regression.
The robust package offers functionality to estimate robust logistic regression models. These are relevant for the same reasons as for robust regression.
