Towards a Better Prescription: Graph AutoEncoder for Drug Recommendation

DOI

10.2991/978-94-6463-654-3_2How to use a DOI?

Keywords

Drug recommender systems; Deep Learning; Graph representation learning; Graph neural networks

Abstract

AI models have been widely used as recommender systems in domains such as nutrition, medicine, health status prediction and physical activity. However, their application to drug recommendation is limited due to the complex nature of the medical data stored in Electronic Health Records (EHRs). Existing AI-based drug recommender models either focus solely on current admission EHRs, neglecting historical records, or learn patient representations separately, ignoring the similarities in medical profiles that could enhance drug prediction accuracy. To address these limitations, we propose a novel Graph Autoencoder Drug Recommender (GADR) framework for predicting drugs for patients based on their similarity scores. First, we model the relationships between patients in an EHR system by creating a population graph, optimally capturing the medical similarities of patients. We then reduce the representation of the population by learning the graph embedding using a graph autoencoder. Finally, we train multiple classifiers using the learned population graph embedding to predict the appropriate drug for a particular patient. Our GADR framework demonstrates promising results on the publicly available MIMIC-III dataset achieving 98% accuracy. This framework is generic and can be applied to other EHR-based 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  - Mongi Kourchid
AU  - Olfa Adouni
AU  - Alaa Bessadok
AU  - Nacim Yanes
PY  - 2025
DA  - 2025/02/24
TI  - Towards a Better Prescription: Graph AutoEncoder for Drug Recommendation
BT  - Proceedings of the  International Conference on Decision Aid and Artificial Intelligence (ICODAI 2024)
PB  - Atlantis Press
SP  - 7
EP  - 22
SN  - 2589-4919
UR  - https://doi.org/10.2991/978-94-6463-654-3_2
DO  - 10.2991/978-94-6463-654-3_2
ID  - Kourchid2025
ER  -