Working Decision Trees and ID3 algorithm!

YAY!! It works!!

I spent today working on getting the navigation of the tree working so that I can start putting it into my OGRE application and start the final stages of my project.

Luckily and to my gratitude, it worked nearly first time with no major errors or mistakes. So right now, the program reads in past examples, calculates the entropy of each of the attributes that make up the examples, and builds a tree based on this. This tree is then navigated until a leaf node is reached at which point the PerfromAction() method is called. This function will no doubt become a lot more complicated as more decision trees are added to the project, but right now, it prints out either “ATTACK!!” or “Don’t attack” based on the outcome of the tree.

This was a big step and because it happened so quickly I can make a start on the graphics app now. The application will be simple to start of with, but will ultimately provide the basis for the final demonstration. It is my aim for this first demonstration that I will have a number of characters in the scene, each representing different tribes, loyalties, and strengths. An intelligent agent will then be added to the scene and will perform an action in front of each of these characters depending on the results of the decision tree. While this is a simple demonstration, it is enough to put across the idea of the final project. This will then be expanded once complete.