KNNDC

K-Nearest Neighbor Classifier for dissimilarity datasets

   [W,K] = KNNDC(D,K)
    W = D*KNNDC(K)
    W = D*KNNDC
    C = S*W

Input
D	Dissimilarity dataset used for training
K	Number of nearest neighbors; if not given, then K is optimized
S	Dissimilarity test dataset. It should be based on the same  representation set as D.

Output
W	Classifier
K	Number of nearest neighbors
C	Classification matrix, dataset

Description

This is the KNN classifier on given dissimilarities. No dissimilarity  space nor embedding is applied.

The classifiers optimizes K in the KNN rule for the given dissimilarity  dataset D. It should be square. If not K=1 will be used. For executing  the classifier on a testset S, formally just K is needed and the  classification matrix C with class confidences is computed by TESTKD.  KNNDC thereby supplies a routine that may be used in a similar way by  routines like CROSSVALD for given dissimilarity datasets.

The classification matrix C can be used for finding labels by LABELD or  for error estimation by TESTC or TESTD.

See also

datasets, mappings, testkd, crossvald, labeld, testc, testd,