Here's the code for a basic decision tree system in Matlab. I'm sharing it in response to a reader comment on the previous note. I re-wrote this system in R, which is what I was planning on sharing, and which I will be later. This code is not as clean as I would typically like to share and Matlab is not open source. 

When you unzip it, you'll see it's split up into two folders, production and release. The production version has a backtesting script wrapped around it so it's probably more interesting, but also slightly more complex. To run the production version use the function "sim2" and for the release version use the function "run". It will download the data from Yahoo, train and run the algorithm, and then give some output. 

The code is actually relatively short and clear so don't get intimidated at first.