Weekly Project Post — Week 7

This week was a big week for working on The NFL Predictive machine. I’ve done a ton of work learning how to add Support Vector Machines into the project and how to use Decision Trees to sort of “predict the predictions” and figure out when the machine makes good predictions and when it makes poor predictions. With that knowledge, it can then ‘nudge’ its expectedly poor predictions more in the right direction. Lots of work, but lots of fun too.

​

This upcoming week I plan on posting the decision trees and naive bayes pages to the website and, if I have time, making the Rapidly Exploring Random Trees page allow the user to click where they want for a start and for a finish. That and perhaps add the unpolished version of the N-Units Away math paper to the website.

​

There are no impediments in the way, as I’m working solo and have all the tools and materials I need to proceed.

​

The process used last week was just a lot of hustle and elbow grease. That and caffeine. Lots more of both will be in use this week.