%STATSKNNC Stats KNN Classifier (Matlab Stats Toolbox)
%
% W = STATSKNNC(A,'PARAM1',val1,'PARAM2',val2,...)
% W = A*STATSKNNC([],'PARAM1',val1,'PARAM2',val2,...)
% D = B*W
%
% INPUT
% A Dataset used for training
% PARAM1 Optional parameter, see CLASSIFICATIONKNN.FIT
% B Dataset used for evaluation
%
% OUTPUT
% W KNN classifier
% D Classification matrix, dataset with posteriors (0-1)
%
% DESCRIPTION
% This is the PRTools interface to the KNN classifier of the Matlab
% Stats toolbox. See there for more information. It is assumed that objects
% labels, feature labels and class priors are included in the dataset A.
% The classification matrix D is for this classifier a 0-1 matrix with just
% a 1 in the column of the assigned class.
%
% SEE ALSO (<a href="http://37steps.com/prtools">PRTools Guide</a>)
% DATASETS, MAPPINGS, CLASSIFICATIONKNN, KNNC
% Copyright: R.P.W. Duin, r.p.w.duin@37steps.com
function W = statsknnc(varargin)
name = 'Stats KNN';
if mapping_task(varargin,'definition')
W = define_mapping(varargin,[],name);
elseif mapping_task(varargin,'training')
A = varargin{1};
data = +A;
labels = getlabels(A);
res = ClassificationKNN.fit(data,labels,varargin{2:end});
W = trained_mapping(A,res);
else % evaluation
[A,W] = deal(varargin{:});
res = getdata(W);
[dummy,post] = predict(res,+A);
W = setdat(A,post,W);
end
return