Narrow Margins: Classification, Margins and Fat Tails

                           Francois Buet-Golfouse 1

              Abstract                man et al., 2000; Schapire et al., 1997) and (Mangasarian,
    It is well-known that, for separable data, the reg-     1999), consider the case of a linear classifier (e.g., logistic
    ularised two-class logistic regression or support      regression or support vector machine) and investigate the
    vector machine re-normalised estimate converges    ...