Initial Use of Hierarchically Optimal Classification Tree Analysis In Medical Research
Paul R. Yarnold · 2013
Univariate optimal discriminant analysis (UniODA) is a non-parametric classification methodology for discriminating between two categories of a binary class (“dependent”) variable—such as sex (male versus female), experimental condition (drug versus placebo), or outcome (dead versus alive)—using a single attribute (independent variable). 1 Hierarchically optimal classification tree analysis (CTA) is a nonparametric, non-linear, multiattribute (i.e., appropriate for use with more than one attribute) classification methodology for discriminating a binary class variable, which is accomplished by chaining together successive UniODA models. 2