k means clustering and SSE

how do I plot SSE for kmeans clustering

Answer 1

[Oleg Komarov]

% Random data
xy = [rand(100,2)*2+3; rand(100,2)];
% Visualize clouds
scatter(xy(:,1),xy(:,2))
close
% Cluster
[idx,c,sse] = kmeans(xy,2);
% Plot clouds with centroids and legend
plot(xy(idx==1,1),xy(idx==1,2),'r.','MarkerSize',12)
hold on
plot(xy(idx==2,1),xy(idx==2,2),'b.','MarkerSize',12)
plot(c(:,1),c(:,2),'kx','MarkerSize',12,'LineWidth',2)
plot(c(:,1),c(:,2),'ko','MarkerSize',12,'LineWidth',2)
legend(['Cluster 1 (SSE: ' num2str(sse(1))],...
       ['Cluster 2 (SSE: ' num2str(sse(2))],...
       'Centroids', 'Location','NW')

[Sai Aparna Kalvapalle]

Sorry but I fail to understand what 'kx','ko' are. Could someone elaborate?

[Farrukh Hassan]

This represents the shape 'ko' the shape of the centroid is circle. and kx shows the shape of the centroid be x.