Attribute-incremental construction of the canonical implication basis

摘要

We propose a new algorithm constructing the canonical implication basis of a formal context. Being incremental, the algorithm processes a single attribute of the context at a single step. Experimental results bear witness to its competitiveness.