How to plot k-means output?

Hello all,

I have some data in 8 text files, I have used 5 of them as my training data and the rest as the testing data. I have classified them using k-means. I would like to visualize the output; in other words, I would like to see the membership of each data point in a graph. Can anybody help me please? Thanks in advance.

Answer 1

[Image Analyst]

How about using scatter and have each class be plotted with different color markers? I think that's the most common way.

[Mnr]

Can you please let me know how I can do that? I am using this command for the classification:[ind c ]=kmeans(Data,3) 'ind' has a dimension of about 90000; I do not know use scatter here. Thanks again.

[Image Analyst]

I don't have the Stats toolbox so I can't use kmeans. So I clustered based on position. Please adapt the demo below. See how I set up a different color depending on what class the data point was assigned to.

x = rand(1000, 1); % 1000 random x,y points
y = rand(1000, 1); % 1000 random x,y points
% Create classes.  Class 1 is if x<y.  Class 2 otherwise
classes = (x < y);
% Make class 1 blue and class 2 red
classColors = zeros(length(x), 3);
markerSizes = zeros(length(x), 1);
for row = 1 : length(x)
  if classes(row)    
    % Class 1 = blue
    classColors(row,:) = [0, 0, 1];
    markerSizes(row) = 10;
  else
    % Class 2 = red
    classColors(row, :) = [1, 0, 0];
    markerSizes(row) = 40;
  end
end
scatter(x,y, markerSizes, classColors, 'LineWidth', 2);
% Enlarge figure to full screen.
set(gcf, 'Units', 'Normalized', 'OuterPosition', [0 0 1 1]);
% Give a name to the title bar.
set(gcf, 'Name', 'Demo by ImageAnalyst', 'NumberTitle', 'Off')

[Fatemeh Shomal Zadeh]

Hi,

I am using your idea and takethis code, but when I am running it, my data are not spread like scatter plot. All of them are in the same line. can you help me regarding that please?

[Image Analyst]

I copied and pasted the code:

x = rand(1000, 1); % 1000 random x,y points
y = rand(1000, 1); % 1000 random x,y points
% Create classes.  Class 1 is if x<y.  Class 2 otherwise
classes = (x < y);
% Make class 1 blue and class 2 red
classColors = zeros(length(x), 3);
markerSizes = zeros(length(x), 1);
for row = 1 : length(x)
  if classes(row)    
    % Class 1 = blue
    classColors(row,:) = [0, 0, 1];
    markerSizes(row) = 10;
  else
    % Class 2 = red
    classColors(row, :) = [1, 0, 0];
    markerSizes(row) = 40;
  end
end
scatter(x,y, markerSizes, classColors, 'LineWidth', 2);
% Enlarge figure to full screen.
set(gcf, 'Units', 'Normalized', 'OuterPosition', [0 0 1 1]);
% Give a name to the title bar.
set(gcf, 'Name', 'Demo by ImageAnalyst', 'NumberTitle', 'Off')

and it gives this image.

Not sure what you did to change it since you forgot to attach your modified code. Attach it with your data and code to read in your data and we'll see if we can fix it. By the way, I do have the stats toolbox now, so I can do kmeans() if needed.