Transformation from Publications to Diabetes Ontology using Topic-based Assertion Discovery

Rohithkumar Nagulapati, Mayanka Chandrashekar, Yugyung Lee

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

During the last decade, we have seen an explosive growth in the number of bio-medical publications. In this paper, we present an Assertion Discovery framework that aims to transform from PubMed publications (diabetes domain) to an ontology, called Diabetes Publication Ontology (DPO). The assertions in the DPO ontology were mapped and integrated with ones in existing diabetes ontologies. The Assertion Discovery framework consists of three main components: (i) Assertion Discovery, (ii) Assertion Alignment, and (iii) Assertion Integration. The proposed approach for ontology generation was based on Stanford CoreNLP for Natural Language Processing, OpenIE (Open Information Extraction) for relation extraction, LDA (Latent Dirichlet Allocation) for topic modeling, and OWL API for ontology generation on the Spark parallel engine. We presented a web-based application for searching diabetes publications as well as retrieving the assertions from the diabetes publications through the DPO ontology.

Original languageEnglish
Title of host publicationProceedings - 2018 IEEE International Conference on Healthcare Informatics Workshops, ICHI-W 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages9-18
Number of pages10
ISBN (Electronic)9781538667774
DOIs
StatePublished - Jul 16 2018
Event6th IEEE International Conference on Healthcare Informatics Workshops, ICHI-W 2018 - New York, United States
Duration: Jun 4 2018Jun 7 2018

Publication series

NameProceedings - 2018 IEEE International Conference on Healthcare Informatics Workshops, ICHI-W 2018

Conference

Conference6th IEEE International Conference on Healthcare Informatics Workshops, ICHI-W 2018
Country/TerritoryUnited States
CityNew York
Period06/4/1806/7/18

Keywords

  • Assertion Discovery
  • Diabetes ontologies
  • Latent Dirichlet Allocation
  • NLP
  • Open Information Extraction
  • PubMed

Fingerprint

Dive into the research topics of 'Transformation from Publications to Diabetes Ontology using Topic-based Assertion Discovery'. Together they form a unique fingerprint.

Cite this