Using Weka for unsupervised clustering

Question

I have data in the following format:

X,Y,sim(X,Y)

That is, a list of triples, with:

• X, the name of an object;
• Y, the name of another object;
• sim(X,Y), a real number expressing the distance between the two objects.

Now, I want to apply some unsupervised clustering algorithm on this data. I had Weka in mind but I would gladly consider alternatives too.

Answer 1

There are plenty of algorithms that can work with similarity matrices:

• Hierarchical Linkage Clustering
• DBSCAN
• OPTICS
• Affinity Propagation
• Spectral Clustering

just to name a few. As for software, I prefer ELKI, it has much much more clustering choices.