At the beginning of this week, I continued to work on implementing PLSA, but I ultimately decided to give it up. If I had more time, I would definitely continue down this path. But simply finding the tools to do it efficiently enough has proved extremely time consuming, and there is plenty of analysis that needs to be done with the other algorithms. I devoted the rest of the week to running each algorithm several more times, gathering more information about each of them, and writing up the results.