GMMB_NORMALIZE Normalize a matrix so that row sums are one. p_out = GMMB_NORMALIZE(p_in) p_in = N x K matrix p_out = N x K matrix If an unnormalized row sum would be zero, the row is left untouched. Author(s): Pekka Paalanen <pekka.paalanen@lut.fi> Copyright: Bayesian Classifier with Gaussian Mixture Model Pdf functionality is Copyright (C) 2003, 2004 by Pekka Paalanen and Joni-Kristian Kamarainen. $Name: $ $Revision: 1.1 $ $Date: 2004/11/02 08:32:22 $
0001 %GMMB_NORMALIZE Normalize a matrix so that row sums are one. 0002 % 0003 % p_out = GMMB_NORMALIZE(p_in) 0004 % 0005 % p_in = N x K matrix 0006 % p_out = N x K matrix 0007 % 0008 % If an unnormalized row sum would be zero, 0009 % the row is left untouched. 0010 % 0011 % Author(s): 0012 % Pekka Paalanen <pekka.paalanen@lut.fi> 0013 % 0014 % Copyright: 0015 % 0016 % Bayesian Classifier with Gaussian Mixture Model Pdf 0017 % functionality is Copyright (C) 2003, 2004 by Pekka Paalanen and 0018 % Joni-Kristian Kamarainen. 0019 % 0020 % $Name: $ $Revision: 1.1 $ $Date: 2004/11/02 08:32:22 $ 0021 % 0022 0023 function p_out = gmmb_normalize(p_in); 0024 0025 K = size(p_in, 2); 0026 divi = sum(p_in, 2); 0027 divi(divi==0) = 1; 0028 0029 p_out = p_in ./ repmat(divi, 1, K);