So today the decision tree took its first gasping and labored breath and lived! What I mean, is that today I finished all the basic guts to make the implementation of my algorithm recursive. (If you don’t know what recursion try searching google for recursion). Recursion is the handy and sometime mind-bending ability of a computer program function, method or object to call/invoke itself from within itself! This ability is very useful because often times an operation that needs to be performed on one block of data is the same as that which is necessary to be performed on each sub-block of data. Therefore the function keeps calling itself until it gets down to the smallest sub-block of code it can process, then after it finishes the next block up the chain uses its results and so on, until the whole thing is wrapped back up at the top. Inherently decision trees (and just about any sort of software tree structure, and even some real trees) can be recursive, because their branches are basically just smaller trees themselves. The decision tree algorithm makes smaller decision trees using recursion until it finds a node with 0 entropy  (which means the outcome is totally determined at that node) called a leaf. In such a way a single object, Decision Tree, can build itself from the leaf nodes back up to the trunk or root node. So far my tree has no fancy output or graphical display and is not commented or ready for public viewing, but as I clean it up and get it interacting with the Arduino I’ll post the polished source. For now here’s a teaser of the command line debug feedback showing the tree at work.