Clustering, algorithms and graphs

Clusters exist everywhere, I try to find them in a particle detector, placed nearly a hundred meters underground in a carven in Geneva. But finding clusters of uhm, stuff.. is principal to many sciences and IT projects these days.

My first clustering algorithm is kind of weird, and wildly inefficient, but have a few interesting treats, perhaps I will document it one day, if it turns out to be any good, at well... anything really...

The study of clustering is part of a project I'm currently doing at my University, and I don't really know where it is heading at the moment, it has kind of diverged into two branches, one where I try to make my own original algorithms (in Python) and one where I simply learn to implement others in C++, CUDA and OpenGL.

Particle Jet finding with C-Means. Size = Momentum, Color = cluster #.

C-Means Clustering Nice convergence, but a few misses.

Cumulative connections Notice the two lower plateaus.

10 links pr node Red marks top 10 most connected.



My Link list cluster algorithm finding something central...

OpenGL viewer Plotting some data in C++ and OGL.