Home > brede > brede_mat_plot_clusterbush.m

brede_mat_plot_clusterbush

PURPOSE ^

brede_mat_plot_clusterbush - Plot a hierarchy of clusters

SYNOPSIS ^

function brede_mat_plot_clusterbush(LL, RR, varargin)

DESCRIPTION ^

 brede_mat_plot_clusterbush - Plot a hierarchy of clusters

       brede_mat_plot_clusterbush(LL, RR)

       Input:    LL   Cell array of 'mat' structures
                 RR   Cell array of 'mat' structures

       Property: FontSize      [ {8} or {11} | positive integer ] Font
                               size of text at the top of nodes. A
                               suitable size may depend on Matlab
                               version and whether the plot is
                               displayed or printed.

                 LineColor     [ {[ 0.8 0.5 1]} | 3 element vector |
                               scalar ] Color of dots specified as RGB
                               values. The values should be between 0
                               and 1. A value of 0.7 will produce 
                               gray lines. 
                 NodeColor     [ {[1 1 0]} | 3 element vector | scalar ]
                               Color of dots specified as RGB
                               values. The values should be between 0
                               and 1. A value of 0.9 will produce a
                               light gray dot.
                 NodeOrdering  [ {Right} | Left | ... ]
                 NodeTextType  [ {Left} | LeftWTA | Right | RightWTA ]
                 NumberOfTexts [ {5} | Integer ] Number of text
                               lines on each node
                 Threshold     [ {0.15} | real ] Threshold for the
                               lines between the nodes

       Construct a 'cluster bush' plot suitable for a hierarchy of
       matrix decompositions. The two input argument may come from,
       e.g.,  the brede_mat_hnmf function.

       'NodeTextType' is determining what text should be displayed on
       top of the nodes. For 'NodeTextType'='Left' the information
       from the rows in the left matrices in LL is used. For
       'NodeTextType'='Right' the information from the columns in the
       right matrices in RR is used. The rows and columns
       corresponding to the largest values in the matrices are
       used. The 'WTA' versions apply a winner-take-all function
       before sorting the elemets of the matrix.
 
       The construction of EPS files from the generated figure may not
       work well if the color EPS is not constructed, i.e., "print
       -depsc"  

       Example:
         froi = fullfile(fileparts(which('brede')), 'xml', 'worois.xml');
         Roi = brede_read_xml(froi, 'output', 'collapsesecond');
         Bs = brede_read_xml('Nielsen2004Database.xml');
         E = brede_bib_bib2exp(Bs{1}.Bib);
         M = brede_exp_exp2mat(E, 'type', 'woroi', 'Roi', Roi);
         M = brede_mat_scale(M, 'type', 'rowzscore'); 
         index = find(sum(~isnan(M.matrix),1)>2);
         M = brede_mat_submatrix(M, 'columns', index);
         [LL,RR] = brede_mat_hkmc(M, 'assignment', 'left')
         brede_mat_plot_clusterbush(LL, RR);
 
       See also BREDE, BREDE_MAT, BREDE_PLOT, BREDE_MAT_HNMF,
                BREDE_MAT_KMC, BREDE_MAT_SUBMATRIX, BREDE_UI_MAT.

 $Id: brede_mat_plot_clusterbush.m,v 1.13 2008/04/10 13:31:45 fn Exp $

CROSS-REFERENCE INFORMATION ^

This function calls: This function is called by:
Generated on Fri 27-Nov-2009 18:11:22 by m2html © 2005