We introduce Formal Concept Analysis and describe a procedure of concept lattice-based ranking, showing how this method is able to install order on objects of retrieval that would otherwise be rendered irrelevant. This is in contrast to best-match retrieval which is prevalent in the field of information retrieval. The strength of the proposed method is its ability to differentiate non-matching items. We conduct experiments to demonstrate the validity of the procedure and investigate its efficiency. In light of the experimental results, we discuss possible refinements to the procedure. We extend the investigation further to data sets in a peer-to-peer setting, making use of a recent algorithm on attribute granularity based on Formal Concept Analysis. We demonstrate with the expe...[ Read more ]

We introduce Formal Concept Analysis and describe a procedure of concept lattice-based ranking, showing how this method is able to install order on objects of retrieval that would otherwise be rendered irrelevant. This is in contrast to best-match retrieval which is prevalent in the field of information retrieval. The strength of the proposed method is its ability to differentiate non-matching items. We conduct experiments to demonstrate the validity of the procedure and investigate its efficiency. In light of the experimental results, we discuss possible refinements to the procedure. We extend the investigation further to data sets in a peer-to-peer setting, making use of a recent algorithm on attribute granularity based on Formal Concept Analysis. We demonstrate with the experimental results the applicability of the procedure and its relations to other methods. Finally, we discuss how our findings are relevant to search engine applications. Our contributions is mainly in demonstrating the utility of Formal Concept Analysis in information retrieval, pointing to possible design reorientation of search systems to take advantage of concept-based ranking.