somtoolbox2/som_neighbors.m (4dbef18) - SVMCrossVal.git

som_neighbors.m

function Ne = som_neighbors(sM,neigh)

% Ne = som_neighbors(sM,neigh)
%
% sM      (struct) map or data struct
%         (matrix) data matrix, size n x dim
% [neigh] (string) 'kNN' or 'Nk' (which is valid for a SOM only)
%                  for example '6NN' or 'N1'
%                  default is '10NN' for a data set and 'N1' for SOM
%
% Ne      (matrix) size n x n, a sparse matrix
%                  indicating the neighbors of each sample by value 1 
%                  (note: the unit itself also has value 0)

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

if isstruct(sM), 
  switch sM.type, 
   case 'som_map',  M = sM.codebook; 
   case 'som_data', M = sM.data; sM = []; 
  end
else
  M = sM; 
  sM = []; 
end

n = size(M,1);

if nargin<2, 
  if isempty(sM), neigh = '10NN'; else neigh = 'N1'; end
end

if strcmp(neigh(end-1:end),'NN'),
  k  = str2num(neigh(1:end-2));
  kmus = som_bmus(M,M,1:k+1);
  Ne = sparse(n,n);
  for i=1:n, Ne(i,kmus(i,:)) = 1; end
else
  if ~isstruct(sM), error('Prototypes must be in a map struct.'); end      
  k  = str2num(neigh(2:end));
  N1 = som_unit_neighs(sM);    
  Ne = sparse(som_neighborhood(N1,k)<=k);
end
Ne([0:n-1]*n+[1:n]) = 0; % remove self from neighbors

return;

SVMCrossVal.git

starting som prediction fine-tuned class-performance visualisation