Knowledge Graphs on Medical Domain Using Mondo Ontology
Chetana B. Thaokar, Madhuri A. Tayal, Pratham K. Gangwal, Rubal S. Singh · 2024
According to the National Library of Medicine, the anatomical disease section includes 18 major categories with more than 26000 known diseases. Although medical domain knowledge is such an important pillar of modern human society, it still consists of a lot of textual medical knowledge. In a few recent years, knowledge graphs emerged as a solution to represent textual data in the form of graphs. In our work, we aim to develop a knowledge graph dedicated toward the medical domain, which will organize textual knowledge into a format that can be easily queried to produce knowledge both faster and in an efficient manner. The source data-set is chosen from Pub-Med, a subsidiary of National Centre of Biotechnology Information, alongside using Mondo ontology for rich structure of knowledge graph. Our framework consists of data preprocessing, entity extraction, relation extraction, triple construction and constructing a data set of RDF triplet. We then used RDF triplet to represent nodes and edges, which we subsequently deployed to Neo4J.