ABSTRACT:
Drug evaluation and safety plays a crucial role in the development and usage of therapeutically effective medications. Traditionally, randomized controlled trials have been the gold standard for assessing drug efficiency and safety. However, these trials often limited number of participants who meet specific eligibility criteria, which may not fully represent the diversity of the target population. Despite ongoing efforts to predict toxicity, accurately forecasting drug side-effects remains difficult. In this research paper we have proposed an approach that leverages side-information sources and compares state-of-the-art machine learning techniques to enhance prediction accuracy. A data analysis pipeline is implemented to obtain relevant side-information for the prediction task. The prediction problem is formulated as a machine learning task to predict side effects for new drugs. We have compared the prediction accuracies of linear and non-linear machine learning methods across ten different side-effects.