This paper aims at reviewing the use of ML algorithms in the identification of DDIs and AEs during drug development. In the past, methods involved did not significantly allow for forecasting DDIs and AEs, largely due to historical data and lack of deeper understanding of pharmacology of drugs. ML algorithms are more efficient due to big data from clinical trials and patients’ EHRs, finding complex relationships that enhance precision. In the paper, it looks at different forms of ML approaches such as regression and neural networks and clustering where a better performance is observed as compared with the conventional approaches. It can be concluded that, the proposed method of using ML for classification demonstrates a higher degree of accuracy, precision and recall compared with the other models and could be impactful in enhancing drug safety and efficacy in future. However, there is still issues such as data quality issues as well as ethical issues managing patient’s data. Future development includes the use of different data and incorporating modern technologies into predicting ability and ethical issues.

Machine Learning, Drug-Drug Interactions, Adverse Events, Pharmacovigilance