[mikedmiked]

Even if the topic is not something you care about the techniques and methods gwern uses are not often talked about here and useful for a variety of text processing tasks.

His methodical analysis and scientific rigor is always what I wish other posts could have. If nothing this is a post which will motivate others to write better.

The cluster analysis is cool - not seen that before. Wonder what the results of Latent Dirichlet Allocation might have gotten...