Changeset 18 for distools/disnorm.m


Ignore:
Timestamp:
06/15/11 15:05:46 (13 years ago)
Author:
bduin
Message:

clevald, parzenddc, parzend_map and kem added

File:
1 edited

Legend:

Unmodified
Added
Removed
  • distools/disnorm.m

    r16 r18  
    11%DISNORM Normalization of a dissimilarity matrix
    22%
    3 %       V = disnorm(D,OPT)
     3%       V = DISNORM(D,OPT)
    44%   F = E*V
    55%
    66% INPUT
    77%   D    NxN dissimilarity matrix or dataset, which sets the norm
    8 %   E    Matrix to be normalized
     8%   E    Matrix to be normalized, e.g. D itself
    99%       OPT  'max' : maximum dissimilarity is set to 1 by global rescaling
    1010%                  'mean': average dissimilarity is set to 1 by global rescaling (default)
     
    1717%   OPT = 'mean'
    1818%
     19% DESCRIPTION
     20% Operation on dissimilarity matrices, like the computation of classifiers
     21% in dissimilarity space, may depend on the scaling of the dissimilarities
     22% (a single scalar for the entire matrix). This routine computes a scaling
     23% for a giving matrix, e.g. a training set and applies it to other
     24% matrices, e.g. the same training set or based on a test set.
    1925
    2026% Copyright: Elzbieta Pekalska, ela.pekalska@googlemail.com
     
    4147%DEFINE mapping
    4248if isstr(opt)
    43         discheck(D);
     49%       discheck(D);
    4450        opt = lower(opt);
    4551        if strcmp(opt,'mean')
Note: See TracChangeset for help on using the changeset viewer.