Histogram difference string distance for enhancing ontology integration in bioinformatics

Integration of bioinformatics ontologies is an important research task. This paper presents a family of new methods of string distance computation for improving existing ontology integration and alignment techniques. A histogram, the main tool of the introduced methods, is an associative array for storing the number of occurrences of each character in a string. We use histogram difference in combination with Longest Common Prefix, TFIDF, Smith-Waterman, and Jaccard re-scorers to define the four members of our family of string matching methods. We compare the performance of our methods with several well-known string matching algorithms using five Gene Ontology datasets as test beds. Our methods outperformed those algorithms in terms of average precision on four datasets and for maximum F₁ measure on three datasets. On the remaining datasets our results were among the best, compared to these well-known methods.