Automatic increment in a knowledge base by means of fuzzy system with supervised machine learning

The integration between Machine Learning (ML) and Fuzzy Systems is a recurring theme in the field of Artificial Intelligence (AI), specially regarding the deductive methods of a Fuzzy System, and, on the other hand, the inductive ones of ML. This article presents an experiment which integrates both approaches, thus showing that they may indeed be complementary. The experiments consists of providing a ML with a mechanism for automatic increment of its knowledge base by means of inserting examples (correctly classified by a Fuzzy Sistem) into supervised learning problems. Increment by means of inserting correctly classified examples allows for a growth of the base and an increase in the ML performance. Finally, in this experiment we show that (under certain conditions), a Fuzzy System ensures the correctness of those examples which will be inserted into the said base and thus ensures an increase in the ML performance.