FIXEDCC

Construction of fixed combiners, back-end routine

    V = FIXEDCC(A,W,TYPE,NAME,PAR)

Input
A	Dataset
W	A set of classifier mappings
TYPE	String defining the type of combination rule
NAME	The name of this combination rule, arbitrary
PAR	Possible parameter for combiner defined by TYPE

Output
V	Mapping

Description

Define a mapping V which applies the combination rule TYPE to the  set of mappings W. The set of mappings W should be a parallel  combination (see MAPPINGS).

TYPE defines the combining rule and can be any of the following
average, min, max, mean, median, prod, vote,

Note that average is only possible for affine 2-class classifiers.

When W is a set of classifiers and A a dataset (possibly the result  of B*W, where W is again a set of classifiers) then

     V = FIXEDCC(W,[],TYPE)   combines the mappings W with the comb. rule TYPE
     V = FIXEDCC(A,[],TYPE)   computes the combining output for dataset A
     V = FIXEDCC(A,W,TYPE)    computes the combining output for dataset A,
                              where W is trained using A

This is a back-end routine. Users should directly call the fixed combiners.

Example(s)

See also

mappings, votec, maxc, meanc, medianc, minc, prodc, averagec,