HODC-AES is intended for research use in computational pharmacovigilance.
The benchmark is derived from spontaneous adverse-event reports. These reports do not prove that a drug combination caused an adverse event. A positive label should be interpreted as a reported signal candidate in the processed benchmark, not as confirmed clinical evidence.
Generated negative samples are contrastive/unobserved examples, not confirmed safe drug-combination adverse-event pairs.
This resource is not medical advice and must not be used for clinical decision-making without independent clinical validation and expert review.
Corresponding author for questions: Asif Elahi (asifelahi488@gmail.com)