Solution to the 1st project of the "Metaheuriscs" graduate course.
To help with the first project, I have prepared for you an example of code base in C++ which contains the basic greedy (CART style) algorithm for a fixed tree depth, as well as some subroutines to load datasets and report solutions. The code is located here: https://github.com/vidalt/Decision-Trees
The first project aims to propose more advanced constructive, local search or metaheuristics for decision tree construction. To evaluate and compare our methods we will use: • A depth limit of 4 • A time limit of 5min per dataset I have included some basic instructions to measure and limit CPU time and fix a random seed (if needed) in your approach.