WhiBo is a framework for design and evaluation of “white-box” component-based decision tree algorithms and their parts. It is intended for use by data mining practitioners, researchers and algorithm developers, but also for teaching of decision tree algorithms. The main idea of WhiBo is to offer standardized decision tree algorithm components which enable simple design and performance testing of decision tree algorithms, and easy extension of the component repository. Currently, WhiBo provides a generic decision tree algorithm, a graphical interface for manual and automatic (evolutionary-based) design and a decision tree component repository. A framework for performance testing of the generic decision tree algorithms is implemented in WhiBo as well.

Documentation on the source code is accessible via the WhiBo web site (www.whibo.fon.bg.ac.rs) as API documentation. The web site also provides installation guide and number of tutorials for end users, algorithm developers, and research scientists.