kcenter_dd.m

模式识别工具箱,希望对大家有用!

%KCENTER_DD k-center data description.
% 
%       W = kcenter_dd(A,fracrej,K)
% 
% Train a k-center method with K prototypes on dataset A.
% 
% See also datasets, mappings, dd_roc

% Copyright: D. Tax, R.P.W. Duin, [e-mail]
% Faculty of Applied Physics, Delft University of Technology
% P.O. Box 5046, 2600 GA Delft, The Netherlands
  
function [W,out] = kcenter_dd(a,fracrej,K,nrtries)

if nargin < 4, nrtries = 25; end
if nargin < 3 | isempty(K), K = 5; end
if nargin < 2 | isempty(fracrej), fracrej = 0.05; end
if nargin < 1 | isempty(a) % empty nndd
  W = mapping(mfilename,{fracrej,N});
  return
end


if isa(fracrej,'double')           %training

  if ~isa(a,'dataset')              %train on training set
    error('I need a dataset to train');
  end
  a = oc_set(a);     % make sure a is an OC dataset
  [nlab,lablist,m,k,c] = dataset(a);

  % train it:
  D = dist(a,a');
  [lab,J,dmin] = kcentres(D,K,nrtries);
  x = a(J,:);

  % obtain the threshold:
  d = min(dist(x,a'))';
  thr = -threshold(-d,fracrej);

  %and save all useful data:
  W.x = +x;
  W.threshold = thr;
  W.scale = mean(d);
  W = mapping(mfilename,W,str2mat('target','outlier'),k,c);

else                               %testing

  [nlab,lablist,m,k,c,p] = dataset(a);
  [W,classlist,type,k,c] = mapping(fracrej);  % unpack

  %compute:
  out = [min(dist(W.x,a'))' ones(m,1)*W.threshold];
  newout = dist2dens(out,W.scale);

  W = dataset(newout,getlab(a),classlist,p,lablist);
end
return