Article Per Year (5 Year)
p-Index From 2020 - 2025
0.23
P-Index
This Author published in this journals
All Journal Journal of Novel Engineering Science and Technology
Agus Rahmat Fadillah
Unknown Affiliation
Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Environmental Science Mechanical Engineering

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Implementation of Knowledge-Based Graph Neural Networks for Reasoning and Ranking Medical Entities from CORD-19 Texts Agus Rahmat Fadillah; Irwansyah Saputra
Journal of Novel Engineering Science and Technology Vol. 4 No. 03 (2025): Forthcoming Issue - Journal of Novel Engineering Science and Technology
Publisher : The Indonesian Institute of Science and Technology Research

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.56741/jnest.v4i02.1049

Abstract

The rapid growth of biomedical literature, espe- cially during the COVID-19 pandemic, has introduced new challenges in retrieving clinically relevant information using conventional search methods. This study proposes a novel, interpretable framework for biomedical information retrieval that integrates Named Entity Recognition (NER), knowledge graph construction, and Graph Neural Networks (GNNs) to support semantic reasoning and entity-level ranking. Unlike prior biomedical retrieval systems that operate at document level or perform link prediction over KGs, our framework introduces a novel task formulation contextual entity-level ranking powered by graph-based semantic reasoning. Leveraging the CORD-19 dataset, the system filters abstracts based on user queries, extracts domain-specific entities using SciSpacy, and constructs a semantic graph that captures co-occurrence relationships among medical concepts. A Graph Convolutional Network (GCN) is then employed to prop- agate relevance signals across the graph, enabling context- aware entity ranking. Experimental evaluations using queries such as ”pneumonia” and ”cough” demonstrate superior performance over traditional IR baselines like TF-IDF and BM25, achieving a Mean Average Precision (MAP) of 0.95 and Precision@3 of 1.00. The results confirm the system’s effectiveness in identifying semantically meaningful biomed- ical entities while offering enhanced transparency through graph-based visualizations. This work contributes a scalable and extensible approach to biomedical search and lays the foundation for intelligent literature exploration in medical research and clinical decision support.
Co-Authors Irwansyah Saputra