Abstract:
The following research aims to create a core biomedical ontology for the domain of
Respiratory Tract Infection. Taking UMLS as a data source. Due to unstructured form of biomedical literature and data available on internet and data repositories, it is hard for researchers
and users to manage that kind of data easily. Ontologies are the best way to create formal
knowledge base of unstructured data in a formal way by considering semantic orientations and
relationship of data available. Ontology development increases recent years as it is easy to
develop expert systems or artificial intelligence based systems using ontologies. In this research
Resource description framework RDF is used for concept mapping and creation of schemas
and based on these schemas respiratory tract infection ontology is enriched with concepts and
attributes like definitions, synonyms, broad and narrow relations within the domain. We use
UMLS as a data source because of its authenticity and broad coverage of bio-medical concepts.
Existing available bio-medical ontologies rather focuses on diseases domain or just
symptoms belongs to a particular disease. There is authentic vocabulary e.g. SNOMED CT and
semantic orientation e.g. UMLS used by many bio-medical ontologies but disease-symptom
links should be available for any disease domain ontology. Disease-symptom relations in a
biomedical ontology needs to be considered for creating a Knowledge base for domain that
will be effective for researchers and users of ontology. In this research the proposed ontology
contains the symptoms linked with their related diseases. The ontology created in this research
named as RTI Ontology that contains hierarchy of diseases concepts, sub-type hierarchy, their
semantic orientation, definitions, relationships with other disease, broad and narrow relations
as well as symptoms. The resultant ontology contain 29 disease concepts with 976 sub-types
of diseases and 52,153 total relation counts. The ontology created is evaluated by domain
experts in term of coverage, completeness and correctness.
