AI-DRIVEN EXPLAINABLE MACHINE LEARNING FOR ADVERSE DRUG REACTION PREDICTION USING GRAPH-BASED PHARMACOVIGILANCE SIGNALS

Predicting adverse drug reactions (ADRs) with explainable AI using Graph NeuralNetworks (GNNs) is what this study offers.The model does a superior prediction performance by integrating heterogeneous pharmacovigilance data and generating interpretable subgraphs.The basis of the risk estimation is patient specific and can unravel ADR causal pathways based on biological and pharmacological mechanisms.