TESTCOST

compute the error using the cost matrix C

  Be Aware! This routine needs a lot of maintenance!! (RD)

    E = TESTCOST(A,W,C,LABLIST)
    E = TESTCOST(A*W,C,LABLIST)
    E = A*W*TESTCOST([],C,LABLIST)

Input
A	Dataset
W	Trained classifier mapping
C	Cost matrix
LABLIST	Labels corresponding to the entries of C

Output
E	Total classification cost

Description

Compute the misclassification cost using the cost matrix C. In  LABLIST the corresponding classes should be listed. Note that  classifier W should be a 'genuine' classifier, in the sense that the  output of the classifier should be transformed using CLASSC. See  for an example the help of costm.m.

See also

testd, costm, setcost, getcost,