matching two points with minimum distance

Question

I'm working on this problem:

There is a 3d binary matrix which i cluster with the kmeans algorithm in matlab;

After that i get a value C which contains the coordinates of the centroids of these clusters in an array, for example:

 C=   30.0000   15.0000   48.5000   

      100.2676   57.7382   80.7489

      57.5000   85.0000   35.0000

      27.5000   50.0000   69.5000

(4 centroids: first one with coordinates (30,15,48,5)) where the rows represent the x,y and z coordinate of each centroid.

When I cluster another matrix I get a second value which also contains the coordinates of the second centroids. now i have to match the centroids from first and second clustering which have minimum distance and create an output where i can see which centroids belong together.

I have tried it with pdist(X) and i get the distances pairwise of the coordinates but i can not match them together..

How to solve this, any idea?

Answer 1

You should use pdist2 instead of pdist.

pdist2 calculates the pairwise distance between two sets of observations, whereas pdist calculates the pairwise distance between all observations in a single matrix.

The documentation of pdist2 specifies the correct syntax and meaning of the output distance matrix:

D = pdist2(X,Y) returns a matrix D containing the Euclidean distances between each pair of observations in the mx-by-n data matrix X and my-by-n data matrix Y. Rows of X and Y correspond to observations, columns correspond to variables. D is an mx-by-my matrix, with the (i,j) entry equal to distance between observation i in X and observation j in Y.

So, you should use this as follows:

D = pdist2(C1, C2);

Note 1: If you have find a function which almost does what you want (i.e. pdist) it is often useful to look into the See Also section of the manual and check if one of these may be helpful.