Daniel@0: function [sTstats,csThist] = som_stats_table(csS,histlabel)
Daniel@0: 
Daniel@0: %SOM_STATS_TABLE Statistics table.
Daniel@0: %  
Daniel@0: % [sTstats,csThist] = som_stats_table(csS)
Daniel@0: % 
Daniel@0: %   sTstats = som_stats_table(csS); 
Daniel@0: %   som_table_print(sTstats);
Daniel@0: %  
Daniel@0: %  Input and output arguments ([]'s are optional): 
Daniel@0: %   csS           (cell array) of statistics structs
Daniel@0: %                 (struct) a statistics struct
Daniel@0: %
Daniel@0: %   sTstats       (struct) a table struct with basic descriptive 
Daniel@0: %                          statistics for each variable
Daniel@0: %   csThist       (cell array) of table structs, with histograms for
Daniel@0: %                          each variable
Daniel@0: %
Daniel@0: % See also  SOM_STATS, SOM_STATS_PLOT, SOM_TABLE_PRINT, SOM_STATS_REPORT.
Daniel@0: 
Daniel@0: % Contributed to SOM Toolbox 2.0, December 31st, 2001 by Juha Vesanto
Daniel@0: % Copyright (c) by Juha Vesanto
Daniel@0: % http://www.cis.hut.fi/projects/somtoolbox/
Daniel@0: 
Daniel@0: % Version 2.0beta juuso 311201
Daniel@0: 
Daniel@0: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Daniel@0: %% arguments
Daniel@0: 
Daniel@0: if isstruct(csS), csS = {csS}; end
Daniel@0: dim = length(csS);
Daniel@0: 
Daniel@0: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%5
Daniel@0: %% action
Daniel@0: 
Daniel@0: sTable = struct('colfmt','','headers',[],'values',[],'span',[]); 
Daniel@0: 
Daniel@0: % summary table of all variables
Daniel@0: sT = sTable; 
Daniel@0: sT.headers = {'name','min','mean','max','std','missing'};
Daniel@0: if ~isnan(csS{1}.nunique), sT.headers{end+1} = 'unique'; end
Daniel@0: %if length(col_values), sT.headers = [sT.headers, col_headers]; end
Daniel@0: sT.values  = cell(dim,length(sT.headers));
Daniel@0: sT.span    = ones([size(sT.values) 2]); 
Daniel@0: 
Daniel@0: %if length(col_values), sT.values(:,end-size(col_values,2)+1:end) = col_values; end 
Daniel@0: %if length(col_spans),  sT.span(:,end-size(col_spans,2)+1:end,:)  = col_spans;  end 
Daniel@0: 
Daniel@0: for i=1:dim, 
Daniel@0:     sT.values{i,1} = csS{i}.name;
Daniel@0:     v = [csS{i}.min,csS{i}.mean,csS{i}.max,csS{i}.std];
Daniel@0:     v = som_denormalize(v,csS{i}.normalization); 
Daniel@0:     vstr = numtostring(v,6);
Daniel@0:     sT.values(i,2:5) = vstr'; 
Daniel@0:     sT.values{i,6} = c_and_p_str(csS{i}.ntotal-csS{i}.nvalid,csS{i}.ntotal); 
Daniel@0:     if ~isnan(csS{1}.nunique),
Daniel@0:         sT.values{i,7} = c_and_p_str(csS{i}.nunique,csS{i}.nvalid);
Daniel@0:     end
Daniel@0: end
Daniel@0: sTstats = sT; 
Daniel@0: 
Daniel@0: % histograms
Daniel@0: csThist = cell(dim,1); 
Daniel@0: for i=1:dim, 
Daniel@0:     sH     = csS{i}.hist; 
Daniel@0:     nvalid = csS{i}.nvalid;
Daniel@0:     nbins  = length(sH.bins); 
Daniel@0:     sT         = sTable; 
Daniel@0:     sT.headers = {[csS{i}.name ' values'],'frequency #','frequency %'};
Daniel@0:     sT.values  = cell(nbins,length(sT.headers));  
Daniel@0:     sT.span    = ones(nbins,length(sT.headers),2);
Daniel@0:     for j=1:nbins,         
Daniel@0:         if length(sH.bins) < csS{i}.nunique, sT.values{j,1} = sH.binlabels2{j};
Daniel@0:         else sT.values{j,1} = sH.binlabels{j}; end
Daniel@0:         sT.values{j,2} = sprintf('%d',round(sH.counts(j)));
Daniel@0:         sT.values{j,3} = p_str(sH.counts(j)/nvalid);
Daniel@0:     end     
Daniel@0:     csThist{i} = sT; 
Daniel@0: end   
Daniel@0:    
Daniel@0: return; 
Daniel@0: 
Daniel@0: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%5
Daniel@0: %% subfunctions
Daniel@0: 
Daniel@0: function vstr = numtostring(v,d)
Daniel@0: 
Daniel@0:     tp = (size(v,2)>1);  
Daniel@0:     if tp, v = v'; end
Daniel@0:     nearzero = (abs(v)/(max(v)-min(v)) < 10.^-d);
Daniel@0:     i1 = find(v > 0 & nearzero); 
Daniel@0:     i2 = find(v < 0 & nearzero);     
Daniel@0:     vstr = strrep(cellstr(num2str(v,d)),' ','');
Daniel@0:     vstr(i1) = {'0.0'};
Daniel@0:     vstr(i2) = {'-0.0'};
Daniel@0:     if tp, vstr = vstr'; end
Daniel@0:     return;
Daniel@0: 
Daniel@0: function str = c_and_p_str(n,m)
Daniel@0:   % return a string of form # (%), e.g. '23 (12%)'
Daniel@0:   if     n==m, p = '100'; 
Daniel@0:   elseif n==0, p = '0';
Daniel@0:   else         p = sprintf('%.2g',100*n/m);
Daniel@0:   end
Daniel@0:   str = sprintf('%d (%s%%)',round(n),p); 
Daniel@0:   return;
Daniel@0: 
Daniel@0: function str = p_str(p)
Daniel@0:   % return a string of form %, e.g. '12%'
Daniel@0:   if round(p*100)>100,  p = sprintf('%3g',100*p); 
Daniel@0:   elseif p==1,          p = '100';
Daniel@0:   elseif abs(p)<eps,    p = '0';
Daniel@0:   else                  p = sprintf('%.2g',100*p);
Daniel@0:   end
Daniel@0:   str = sprintf('%s%%',p); 
Daniel@0:   return;