NNERROR1

Exact expected NN error from a dissimilarity matrix (1)

    E = NNERROR1(D,M)
    E = D*NNERROR1([],M)

Input
D	NxN dissimilarity dataset
M	Vector with desired number of objects to be selected

Output
E	Expected NN errror

Description

Exact computation of the expected 1-nearest neighbor error. D should be a  dataset containing a labeled square dissimilarity matrix of a set of  objects related to itself. It is not needed that D is symmetric.  So NNERROR1(D) ~= NNERROR1(D') is possible.

    E = NNERROR1(D)
    E = D*NNERROR1

In this case a set of training set sizes is used to produce a full  learning curve. E can be plotted by PLOTE.

There is a similar routine NNERROR2 which is based on the selection of an  equal number of objects per class.

See also

datasets, nnerror2, nne, testkd,