An optimized human diseases ontology for medical diagnosis

Abstract

The Disease Ontology has more than 8685 classes that define many kinds of diseases, but it is lack of relations of the disease with symptoms and other disease properties (consequences, agents, etc.). Currently, there are also many attempts to organize this information, such as: Diseases Ontology (DOID), Symptoms Ontology (SYMP), and Diseases Symptoms Ontology (DSO) [1]. However, these ontologies are established in isolation, it presents the disease definition with few relations of the disease and disease properties, and it is lack of relations of the disease and other disease properties. In this thesis, we are proposing a method of constructing a Human Disease Ontology for Medical Diagnosis (HDO-MD) by transforming data in the Disease Ontology. The constructed ontology is composed of not only the disease definition and relations of the disease and symptoms, but also the relations of the disease and other disease properties (consequences, agents, etc.). The constructed ontology is implemented and evaluated. Keywords: Transform Disease Ontology, HDO-MD, DOID, SYMP, DSO