Evolutionary Learning of Decision Trees

Software implemented by Marek Grzes
Provides algorithms for learning various types of classification trees using evolutionary search. Includes decision trees with single and multi-attribute tests, cost-sensitive learning, and provides an extensive framework for running and comparing classification algorithms. Runs cross-validation automatically and outputs gnuplot files when several algorithms are evaluated
Written in portable C++, compiles in Linux and Windows in Visual Studio, reads data in the C4.5 input format
Download the source code

My earlier publications on decision trees are based on this code
I am not developing or extending this software any more. If you are interested in extensions that contain more algorithms please refer to work of Prof. Marek Kretowski