Week 3

Class:
Things are starting to get a bit more tricky when it comes to quizzes on Python. We are now getting into depth on how certain objects in python behave, especially when they are iterated over. Before this week I never used to look over notes from the previous class to do well on the quizzes but it is now becoming almost a necessity as Im starting to miss a couple of points here and there. It shouldn't be anything hard to recover from but I need to stay on top of it. The exam is coming up soon and I feel it might be time to start looking over notes from the past two weeks to make sure nothing catches me off guard.

Project:
The Netflix challenge was a lot of fun. Not only was it relevant to real life assignments where a company might ask you to generate a rating system for their products but it was also an opportunity to take advantage of python's speed and flexibility at scripting. I learned more from writing scripts to generate caches than I did from actually trying to beat the Netflix RMSE. When my partner and I started the project we felt really good about our ideas on how to get below an RMSE of 1.00. The obvious idea was to calculate the mean rating for every movie and the mean rating that a given customer would give out. Given the movie ID and the customer ID, one could simply query a cache and take the average of the two values. This got us down to 1.001 which was just out of reach for getting a non-zero grade. We then calculated the medians instead of the means and used a similar approach which got us to just under one. But we didn't stop there. The paper listed on the project page was the key to getting under 1.00 and based on the fact that the professor never really stressed reading it, it is my assumption that he wasn't counting on many of us reading it. Regardless, that paper outlined how to get under 1.0 and if you caught the professors hint about using averages from specific decades then you could easily get the bonus points as well which we were able to do.

On to week 4 :)