showresults.m

这个是支持向量聚类机用matlab编译的主程序和部分子程序

%=====================================================================
%
%   ShowResults
%   ----------------
%
%   Parameters:   
%		Samples         - This matrix holds the data points.
%		SV              - A Matrix containing the Support vectors.
%		BSV             - A Matrix containing the outliers.
%		R		        - The minimal sphere's radius.
%		beta 	        - The vector of the Lagrangian multipliers.
%		quad	        - The quadratic part of the distace from the sphere's
%					      center.
%       q		        - The width of the gaussian kernel.
%		clusters_assignments -
%		        A vector of the clusters assignments assigned by the algorithm 
%		        to the data points.
%       grids           - Each layer in this matrix holds a grid mapping for 
%                       - a couple of dimensions.
%       grids_sizes     - The number of points in each grid layer.
%       classifications - The apriori classifications for each Data point.
%       nof_samples_per_class_per_cluster
%
%	Shows graphs of the aquired results.
%
%=====================================================================

function ShowResults(Samples,SV,BSV,R,beta,quad,q,clusters_assignments,grids, grids_sizes,classifications,nof_samples_per_class_per_cluster);

[attr,N] = size(Samples);


 
grid_index = 0;
for attr1 = 1:attr-1
    for attr2 = attr1+1:attr
        
        % Plot the data points, highlights SVs and BSVs
        figure;
        plot(Samples(attr1,:),Samples(attr2,:),'bd',BSV(attr1,:),BSV(attr2,:),'r*',SV(attr1,:),SV(attr2,:),'g*');
        title(strcat('The Sample Points and the Support Vectors - Dimensions:',num2str(attr1),' and:',num2str(attr2))); 
        xlabel(strcat('Attribute: ',num2str(attr1))); 
        ylabel(strcat('Attribute: ',num2str(attr2))); 
        %legend ('Samples', 'Bounded Suuport Vectors' ,'Support Vectors',-1);

        % show the 2d mapping of the clusters 
        grid_index = grid_index+1;
        figure;
        plot(grids(1:grids_sizes(grid_index),1,grid_index),grids(1:grids_sizes(grid_index),2,grid_index),'k.',Samples(attr1,:),Samples(attr2,:),'bd');
        title(strcat('The Sphere Mapping into a 2D space - Dimension:',num2str(attr1),' and:',num2str(attr2))); 
        xlabel(strcat('Attribute: ',num2str(attr1))); 
        ylabel(strcat('Attribute: ',num2str(attr2))); 
        %legend ('Grid Points','Samples',-1);

        % shows the different clusters 
        ShowClusters([Samples(attr1,:);Samples(attr2,:)],clusters_assignments,attr1,attr2);
    end

end

unique_classifications = unique(classifications);
figure;
bar(nof_samples_per_class_per_cluster,'stacked');
title('Classfications Per Class');
xlabel('Clusters');
ylabel('Number of Samples of each class');
% create the legend string
legend_str = 'legend('
for class = 1:length(unique_classifications);
   if(class ~= 1)
      legend_str = strcat(legend_str,',');
   end   
   legend_str = strcat(legend_str, '''class:',num2str(unique_classifications(class)),'''');
end
legend_str = strcat(legend_str,');');
eval(legend_str);