%% Template file for Iris Data Clustering

%Load the data:

load IrisNoLabel
[mm,nn]=size(X);  % X is number of points by dimension- in this case, 150 x 4

% Initialize the clusters as matrix C, 10 x 4 (Using random data points)

% Run the LBG Update algorithm 10 times (shouldn't need to change this):
for j=1:10
    [P,C,Err]=lbgUpdate(X,C);
end

%Plot the data and the clusters in the first two dimensions (No changes
%here):
figure(1)
plot(X(idx1,1),X(idx1,2),'k.',X(idx2,1),X(idx2,2),'k*',X(idx3,1),X(idx3,2),'k^');
hold on
plot(C(:,1),C(:,2),'rs');
hold off

%%  Visualize the Data

% Here, we want to visualize the data and the cluster centers using the
% best basis.  Find the best basis for the data, and project both the data
% and the cluster centers to the best plane.  End result should be two
% matrices, CoordsX (150 x 2) and CoordsC (10 x 2):



% Plotting routines- Should not need to change anything here:

figure(2)
voronoi(CoordsC(:,1),CoordsC(:,2));
hold on
plot(CoordsC(:,1),CoordsC(:,2),'rs');
plot(CoordsX(idx1,1),CoordsX(idx1,2),'k.',CoordsX(idx2,1),CoordsX(idx2,2),'k*',CoordsX(idx3,1),CoordsX(idx3,2),'k^');
hold off

% Comment on how well the clustering did: