The Complete SPRLIB & ANNLIB

calc_multiclass_error

- calculate the error in classification using the Mahalanobis distance

SYNOPSIS

double calc_multiclass_error (lset, tset, trace, function)

ARGUMENTS

DATASET *lset The learning set.
DATASET *tset The test set.
int trace If trace = 1 produce intermediate results.
int function If function = 1, use the learning set for error calculation, otherwise use the test set.

RETURNS

If no error occurs the function returns a positive number indicating the classification made on either the learning set or test set. In case of errors the function returns a negative number.

FUNCTION

This function computes the classification error based on the Mahalanobis distance. It receives a learning set (lset) and a test set (tset). Depending on function it either computes the error on the learning set (function = 1) or on the test set (function = 0). The learning set is always used to compute the coefficients. If desired intermediate results can be displayed by setting trace to 1.

SEE ALSO

Calc_Dist_Class, mahal_multiclass_dataset, mahal_dataset

This document was generated using api2html on Thu Mar 5 09:00:00 MET DST 1998