Abstract :
We propose to discuss the role of granular computing in tasks of machine learning, knowledge discovery, and data analysis. In the realm of decision systems, we introduce granulation mechanisms which lead to granular structures made into decision systems again. Granulation schemes depend on logical inference mechanisms known from many-valued, in particular fuzzy, logics. On the basis of granulated decision systems, classifiers are constructed in order to test the effectiveness of the given logic in approaching the inner decision logic of the decision system. The results witness that this approach is effective and that granulation does not cause loss of information contained in original
Keywords :
data mining; fuzzy logic; inference mechanisms; learning (artificial intelligence); data analysis; decision systems; fuzzy logic; granular computing; inner logics; knowledge discovery; logical inference mechanisms; machine learning; Aggregates; Data analysis; Fuzzy logic; Fuzzy set theory; Inference mechanisms; Logic testing; Machine learning; Set theory; System testing; Voting;
