Hybrid Graph Attention Network and XGBoost Framework for Interpretable Molecular Toxicity Prediction
Abstract
Traditional experimental toxicity testing is costly, time-consuming, and raises ethical concerns, creating an urgent need for computational alternatives that can rapidly screen chemical safety while maintaining interpretability for regulatory acceptance.
Fields of Study & Subjects
Computational Drug Discovery Methods
Machine Learning in Materials Science
History and advancements in chemistry
Publication details
- Journal / source
- International Journal of Scientific and Research Publications
- Volume
- 16
- Issue
- 1
- Pages
- 223–247
- Publication date
- 24 January 2026
- Language
- EN