DiscoverPath: A Knowledge Refinement and Retrieval System for Interdisciplinarity on Biomedical Research.

Area of Science:

• Biomedical Research
• Information Science
• Computer Science

Background:

• Scholarly publications are growing exponentially, making efficient retrieval difficult.
• Traditional search engines struggle with interdisciplinary research and diverse terminologies.
• A need exists for advanced tools to aid researchers in discovering relevant articles.

Purpose of the Study:

• To present DiscoverPath, a knowledge graph-based search engine for biomedical research.
• To enhance user experience in discovering relevant queries and articles.
• To foster interdisciplinary knowledge exploration.

Main Methods:

• Utilized Named Entity Recognition (NER) and part-of-speech (POS) tagging to extract terms and relationships from abstracts.
• Constructed a knowledge graph (KG) from extracted information.
• Developed a focused subgraph visualization and a query recommendation system.
• Integrated an accessible Graphical User Interface (GUI) for intuitive interaction.

Main Results:

• DiscoverPath effectively extracts terminologies and relationships to build a KG.
• The system reduces information overload by presenting focused subgraphs.
• Query recommendations enable iterative refinement of search queries.
• The GUI provides intuitive visualization and detailed article information.

Conclusions:

• DiscoverPath enhances biomedical research article retrieval through KG-based search.
• The system facilitates interdisciplinary knowledge exploration by improving query and article discovery.
• DiscoverPath offers an accessible and efficient solution for navigating complex scientific literature.