After discussing the implementation of TClass, it is now applied to several datasets to evaluate its behaviour.
This chapter begins with a discussion of what the issues to be investigated are and how the investigation will proceed.
TClass is tested on two artificial domains: the cylinder-bell-funnel domain, and a new artificial classification problem, the TTest domain. Using the lessons learnt on the artificial datasets, two real datasets are then tested: Auslan sign recognition and ECG analysis.