CNORMC

Classifier normalisation for ML posterior probabilities

    W = CNORMC(W,A)

Input
W	Classifier mapping
A	Labeled dataset

Output
W	Scaled classifier mapping

Description

The mapping W is scaled such that the likelihood of the posterior  probabilities of the samples in A, estimated by A*W*SIGM, are maximized.  This is particularly suitable for two-class discriminants. To obtain  consistent classifiers in PRTools, it is necessary to call CNORMC in the  construction of all classifiers that output distances instead of densities  or posterior probability estimates.

If A has soft labels or target labels, W is returned without change.

See also

mappings, datasets, classc,