Background knowledge in formal concept analysis: constraints via closure operators

The aim of this short paper is to present a general method of using background knowledge to impose constraints in conceptual clustering of object-attribute relational data. The proposed method uses the background knowledge to extract only particular clusters from the input data—those which are compatible with the background knowledge and thus satisfy the constraint. As a result, the method allows for extracting less clusters in a shorter time which are in addition more interesting. The paper presents the idea of constraints formalized by means of closure operators and introduces such constraints to a particular clustering technique, namely to formal concept analysis. Among the benefits of the presented approach are its versatility (the approach covers several examples studied before, e.g. extraction of closed frequent itemsets in generation of non-redundant association rules) and computational efficiency (polynomial time-delay algorithm for computing constrained clusters). Due to s...

• Type: Conference
• Year: 2010
• Where: SAC
• Authors: Radim Belohlávek, Vilém Vychodil
• Keywords: Applied Computing, Background Knowledge, Formal Concept Analysis, Particular Clustering Technique, SAC 2010