Conformal Evaluator provides a statistical evaluation of a broad class of machine learning algorithms; such an evaluation builds the core mechanism to enable the identification of concept drift in learning models.

For further information on the project, source code and dataset please send an email to Lorenzo Cavallaro.

Feel free to check out our USENIX Security 2017 presentation, paper, and talk.

Conformal Evaluator is partially yet generously sponsored by the UK EPSRC (EP/L022710/1 and EP/K033344/1)