Fuzzy-rough-learn 0.2: a Python library for fuzzy rough set algorithms and one-class classification
Oliver Urs Lenz, Chris Cornelis, Daniel Peralta · 2022 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE) · 2022
We have expanded the scope of fuzzy-rough-learn 0.2 from fuzzy rough sets to also cover one-class classification, facilitating the exploration of practical and conceptual connections between these two areas of machine learning. The new algorithms for one-class classification consist of nine data descriptors and one feature preprocessor. In addition, we have added Fuzzy Rough Nearest Neighbour regression and a number of preprocessors for feature scaling. Apart from these new core algorithms, we have rewritten fuzzy-rough-learn from the ground up, and included a large number of utility functions, to allow users to easily adapt any of the algorithms to their use case.