depth first search

Multidimensional scaling is next up in my continuing series of Python implementations of famous algorithms. Suppose you have the interpoint distances between some set of points and you want to plot some set of points (preferably 2-D) that have approximately the same interpoint distances. Well, multidimensional scaling provides just such a method.

You can find my implementation here.

Here is an array of 100 points:

Here is are the points generated by mds applied to the interpoint distances:

As you can see, mds recovers the original points up to some isometry.