Toward the inner logic in data: Granular structures as the frame for classifiers

Any analysis of data collected in data tables aims at revealing the inner logic of data, i.e., the mechanism of inference which decides about assignment of classes or categories to single items in data. We propose to discuss and analyze the mechanism based on granulation of data and the subsequent induction of classifiers from the resulting granular structures. We consider some logical operators transferred into the realm of data tables and made a basis for granulation heuristics along with classification results inferred from the resulting granular structures.