Short Letter on the Semiparametric Transformation Discriminant Analysis
Jinghao Xue and Mike Titterington
technical report 2007.

Abstract

The purpose of this letter is to comment that, for the semiparametric transformation discriminant analysis~\citep{Lin:2003}, the approximate method-of-moments estimators or approximate maximum likelihood estimators are not necessary to estimate the mean, and for the linear discriminant case the variance, of the transformed negative class.