Prediction of Drug Reactions through Hybrid Techniques

Everyday, doctors discuss drug prescriptions with different patients. Therefore, while prescribing drugs, doctors must be aware of all potential side effects of drug. According to the US Department of Health and Human Services, 33 percent of all hospital admissions each year were due to Adverse Drug Events (ADEs), also known as unwanted side effects. The goal of this research is to identify adverse reaction of drugs based on user’s reviews. Two techniques are proposed in this research named as Support Vector Machine (SVM) and Extreme Learning Machine (ELM). Further the hyper-parameters of proposed techniques are dynamically optimized through Random Search (RS) and Elephant Harding Optimization (EHO) techniques respectively. The performance evaluation matrices are used to obtain results of proposed techniques. Furthermore, proposed hybrid techniques ELM-EHO and SVM-RS are compared with some conventional techniques to prove superiority.