How can I measure the mean of the Euclidean distances to each point in a scatterplot?

Hi everyone!

I'm new in the Matlab word and I need some help to do some of my result (master student).

I have a scatter plot, and I need to measure the mean of the Euclidean distances to each point to the nearest neighbor in the bi-plot space.I found some information about the knnsearch function, but I realy don't know how to use it.

Thanks

Answer 1

[Image Analyst]

If you have the Stats Toolbox, you can use pdist2(), min(), and mean():

x = rand(10,1);
y = rand(length(x), 1);
plot(x, y, 'b*');
grid on;
% Get distances of each point to every other point.
D = pdist2(x, y)
% Find the min distance for each point
% That is the distance to the closest neighbor.
minDistances = min(D, 2)
% Get the mean of those closest distances
meanClosestDistance = mean(minDistances)

If you don't have that toolbox, you can easily calculate D with a double for loop.

[Noemie Friscourt]

Thank you !

Answer 2

[Star Strider]

The ‘knn’ functions are used for classification. If you just want the distances, use the pdist funciton, and if you want to know what the distances are between various points, use the squareform function to get the matrix.

A = randi(99, 15, 2);                           % Create Data (15x2) Matrix
D = pdist(A, 'euclidean');                      % Calculate Distances
S = squareform(D);                              % Matrix Depiction (Optional)

[Noemie Friscourt]

Thank you for your help !

[Star Strider]

My pleasure!