Ontology-based knowledge extraction from scientific biomedical texts

With the increase in rates of scientific biomedical literature, manual extraction and compilation of knowledge is becoming impossible to accomplish. Ontology-based techniques present a systematic way of mechanically extracting and describing bio-medical information present in free-form texts. In this paper, an expansion is featured regarding the designing and application of an ontology based model of extracting knowledge in the biomedical literature. Based on a combination of natural language processing (NLP), named entity recognition (NER), semantic reasoning, and domain-specific ontology such as the Gene Ontology (GO) and Unified Medical Language System (UMLS), the proposed method produces structured information based on textual information. The framework was tested on the set of PubMed papers with a high accuracy rate in the mapping of terms and extraction of relations. The implications of the findings indicate that semantic technologies will play an important role in the interpretability and usability of biomedical data to downstream applications including decision support and biomedical discovery.