Li near r egr essi on, acti ve learning We arriv ed at the lo gistic regression model when trying to explicitly model the uncertainty about the lab els in a linear c la ss ifier. The same genera l modeling approach p e rmits us to use line a