Designing a Deep Learning Model for recommendation of Drugs on Sentiment Analysis

There has been a rise in creative automation work in recent years, and various applications have been found for machine learning. This essay's focus is on introducing a pharmaceutical recommendation system with the potential to greatly reduce the burden on experts. With the help of a deep learning algorithm, we create a drug recommendation system that utilises patient reviews to predict sentiment through the use of a number of vectorization techniques. Synthetic minority over-sampling method (SMOTE) is used to rectify the effects of unbalanced data on classification accuracy. The evaluation is sorted using a Recurrent Neural Network (RNN), with the weights chosen using an enhanced version of the firefly optimisation algorithm (IFFOA). Predicted feelings were evaluated using a battery of metrics counting precision, recall, f1score, accuracy, and specificity score. The classifier RNN-FCR has a higher accuracy rate (96%) than any other models as a direct result of TF-IDF vectorization.