Adversarial Graph Neural Network for Medication Recommendation (AGMR)

DOI

10.2991/978-94-6463-654-3_5How to use a DOI?

Keywords

Healthcare Recommendation; Drug Recommendation; Graph Representation Learning; Graph Generative Adversarial Network

Abstract

Deep learning has transformed recommender systems by enabling them to thoroughly analyze and comprehend intricate patterns in user behavior and preferences. With the ability to handle large volumes of data, these systems can make accurate predictions and offer personalized suggestions in the healthcare sector. However, deep learning models for medication recommendation face challenges in addressing diverse patient demographics and integrating complex data, leading to biased and inaccurate suggestions. The process also overlooks certain medications and uses a uniform number of hops for all patients, despite varying amounts of information in their electronic health records. To address these challenges, we propose the Adversarial Graph Neural Network for Medication Recommendation (AGMR). This model is designed with two primary components. The first component involves constructing a population graph to analyze the complex relationships between patients in an EHR system, effectively capturing their medical similarities. We then use graph GANs to learn and reduce the graph representation. The second component involves training various classifiers on the generated population graph embedding to predict the most appropriate medication for each patient. The AGMR has demonstrated impressive results on the available MIMIC-III dataset, achieving an accuracy of 0.88%. This generic framework can also be applied to other EHR-related datasets.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Oussama Abdeddaiem
AU  - Atta Zaatra
AU  - Alaa Bessodok
AU  - Nacim Yanes
PY  - 2025
DA  - 2025/02/24
TI  - Adversarial Graph Neural Network for Medication Recommendation (AGMR)
BT  - Proceedings of the  International Conference on Decision Aid and Artificial Intelligence (ICODAI 2024)
PB  - Atlantis Press
SP  - 47
EP  - 61
SN  - 2589-4919
UR  - https://doi.org/10.2991/978-94-6463-654-3_5
DO  - 10.2991/978-94-6463-654-3_5
ID  - Abdeddaiem2025
ER  -