Adverse Drug Reactions (ADRs) are mild-to-serious undesired effects a patient can undergo after taking a particular medication. ADRs can even be life threatening for certain subpopulations of patients. Overall, ADRs present a major financial burden to the US healthcare system . One important challenge for accurately monitoring both the occurrences and financial burdens associated with ADRs is mainly the lack of a cohesive definition used to track such events . According to the FDA Adverse Event Reporting System (FAERS), there were over 1.2 million adverse events reported in 2014 compared to only 400,000 cases in 2006 . For all the aforementioned reasons, there is a strong need to develop computational models that could accurately predict a drug's likelihood of causing ADR. In fact, models reliably predicting such ADR for a subpopulation of patients are especially in high demand for future precision medicine protocols.
There are two main different types of ADRs referred to as Type I and Type II [4, 5]. Type I ADRs are predictable pharmacological events induced by a drug, ...