Attribute Selection Using Contranominal Scales

Formal Concept Analysis (FCA) allows to analyze binary data by deriving concepts and ordering them in lattices. One of the main goals FCA is enable humans comprehend information that encapsulated data; however, large size concept lattices a limiting factor for feasibility understanding underlying structural properties. The such lattice depends on number subcontexts corresponding formal context are isomorphic contranominal scale high dimension. In this work, we propose algorithm ContraFinder enables computation all scales given context. Leveraging algorithm, introduce \(\delta \)-adjusting, novel approach order decrease selection an appropriate attribute subset. We demonstrate \)-adjusting reduces hereby emerging sub-semilattice implication set restricted meaningful implications. This evaluated with respect its associated knowledge means classification task. Hence, our proposed technique strongly improves understandability while preserving important conceptual structures.