Multiple views for ontology-based formal concept lattices

Formal concept analysis (FCA) has developed into an accepted knowledge discovery paradigm being productively used in solving practical problems involving clustering, classification or association discovery. FCA is being applied to annotation files and their associated ontological terminology. Here the formal context is a matrix whose rows correspond to the annotated objects and whose columns correspond to the annotation terms that exist in an ontological terminology. This paper describes a software tool Multi-FCA that generates multiple views of the formal context. A view is determined by the emphasis on objects, properties, or complementation. In addition, these views can be filtered using the structure of the ontological terminology. A simple example using a small portion of the Gene Ontology (GO) is used to illustrate the filtered view for a concept lattice. Future work to improve the efficiency of reconstructing a filtered view and incorporate the degree of evidence supporting the gene-annotation mapping is described.