Cluster and rank blogs by logical categories

Question

What kind of algorithm would be good to cluster and rank blogs in logical communities (tech, entertainment, etc...)?

An algorithm to cluster and rank blog posts would be even better.

Answers accepted are algorithms, pseudo-code, java code or links to explanations on particular algorithms.

Update: So, it seems I would like something in the category of Partional Clustering based, mostly, on textual features.

Answer 1

First you have to define your problem better by defining a:

1. Feature vector (a descriptor) for each blog
2. A metric to evaluate the "distance" between descriptors.

Once you've done that, you have a truckload of clustering algorithms to choose from, such as kmeans, etc.