A PREDICTIVE MODEL FOR DRUG-DRUG INTERACTION (DDI) USING MACHINE LEARNING TECHNIQUES
Nlerum Promise Anebo, Samuel Apigi Ikirigo · International Journal of Machine Intelligence · 2025
This study focuses on developing a robust and accurate drug-drug interaction (DDI) predictive model, leveraging advanced machine learning techniques, specifically Random Forest and XGBoost.The core objective is to create a comprehensive polypharmacy DDI detection tool primarily aimed at identifying potential and harmful interactions among diabetic patients, a demographic particularly vulnerable due to complex and often extensive medication regimens.Beyond mere detection, the system is meticulously designed to provide a highly intuitive and user-friendly web-based interface, making critical DDI information readily accessible to healthcare professionals and patients alike.This accessibility is paramount in fostering informed medication management decisions and, ultimately, enhancing overall patient safety and promoting greater adherence to prescribed treatment plans.The hybrid model developed through this research demonstrates exceptional predictive capabilities, achieving a high accuracy of 98.45% and an impressive F1 score of 0.9840, signifying its strong performance in correctly identifying true interactions while minimizing false A Predictive Model for Drug-Drug Interaction (DDI) using Machine Learning Techniques https://iaeme.com/Home/journal/IJMI 2 [email protected] and negatives.This robust performance positions the model as a valuable asset in mitigating the risks associated with polypharmacy in diabetic care, especially within resource-constrained environments.