A fuzzy set framework for ontological similarity measures

Advances in biotechnology have given rise to rapid production of biomedical data and the creation of a wide variety of biomedical and bioinformatics ontologies which serve as a primary impetus for the creation of new ontological similarity measures. A new IC ontological similarity measure, a variation of a standard fuzzy set similarity measure, is presented. This new IC ontological similarity measure has never been used in bioinformatics studies. It is compared with two standard IC ontological similarity measures, Resnik and Lin, and also an existing modification of Lin´s measure by analyzing the results when used on a set of concepts pairs selected from the cellular component sub-ontology of the Gene Ontology. As part of this experiment, several different methods of calculating IC for an ontological concept are also investigated. These results for the cellular component sub-ontology confirm the results of a previous study on WordNet showing that ontology-based IC measures corresponded closely with corpus-based IC measures.