

   CCoonnddeennssee ttrraaiinniinngg sseett ffoorr kk--NNNN ccllaassssiiffiieerr

        condense(train, class, store)

   AArrgguummeennttss::

      train: matrix for training set

      class: vector of classifications for test set

      store: initial store set. Default one randomly chosen
             element of the set.

   DDeessccrriippttiioonn::

        Condense training set for k-NN classifier

   DDeettaaiillss::

        The store set is used to 1-NN classify the rest, and
        misclassified patterns are added to the store set. The
        whole set is checked until no additions occur.

   VVaalluuee::

        index vector of cases to be retained (the final store
        set).

   RReeffeerreenncceess::

        P. A. Devijver and J. Kittler (1982) Pattern Recogni-
        tion. A Statistical Approach.  Prentice-Hall, pp.
        119-121.

   SSeeee AAllssoo::

        `reduce.nn', `multiedit'

   EExxaammpplleess::

        data(iris3)
        train <- rbind(iris3[1:25,,1],iris3[1:25,,2],iris3[1:25,,3])
        test <- rbind(iris3[26:50,,1],iris3[26:50,,2],iris3[26:50,,3])
        cl <- factor(c(rep("s",25),rep("c",25), rep("v",25)))
        keep <- condense(train, cl)
        knn(train[keep,], test, cl[keep])
        keep2 <- reduce.nn(train, keep, cl)
        knn(train[keep2,], test, cl[keep2])

