wolffd@0: function w = svmlight2weight(modelfile)
wolffd@0: % returns the dual weight vecor for a given svm model
wolffd@0: %
wolffd@0: % All the script does is compute the weighted sum of the support vectors
wolffd@0: % (first element in line is alpha*y, what follows is the feature vector).
wolffd@0: % For further info, see the comment in the model file for its format. 
wolffd@0: 
wolffd@0: tagstr = python('svmlight2weight.py', modelfile);
wolffd@0: 
wolffd@0: [firstidx] = strfind(tagstr, '1 :');
wolffd@0: 
wolffd@0: %get start of vector description
wolffd@0: tagstr = tagstr(firstidx:end);
wolffd@0: 
wolffd@0: % parse text function output
wolffd@0: vecs = textscan(tagstr,'%d %f','Delimiter',':');
wolffd@0: 
wolffd@0: % initialise
wolffd@0: w = zeros(max(vecs{1}),1);
wolffd@0: 
wolffd@0: % set values
wolffd@0: w(vecs{1}) = vecs{2};
wolffd@0: