DRUG RECOMMENDATION SYSTEM BASED ON SENTIMENT ANALYSIS OF DRUG REVIEWS USING MACHINE LEARNING

Abstract

Due to issues such an absence of qualified medical staff, insufficient hospital infrastructure, and essential drugs and medical equipment, the accessibility of high-quality healthcare resources has been steadily decreasing since their establishment. Consequently, individuals all over the globe are placing a heavy burden on healthcare systems and, in some cases, endangering their own health by attempting to self-medicate.A lot of people are interested in machine learning because of its potential to help with healthcare decision-making by automating and augmenting it. This study proposes a medication recommendation system to assist patients in selecting appropriate medications, with the aim of facilitating the work of healthcare professionals. The system assesses sentiment and makes predictions based on patient medication ratings using a host of text representation algorithms, such as Bag of Words (BoW), TF-IDF, Word2Vec, and proprietary feature extraction methods.A number of classification methods are used to assess the processed textual input and to forecast its emotion. The models are evaluated using common evaluation metrics such as recall, accuracy, precision, F1-score, and AUC score. Combining the Linear Support Vector Classifier (LinearSVC) with TF-IDF vectorisation yields the best performance, according to the experimental results. This strategy outperforms all others that were considered. When used together, these factors provide an accuracy of around 93%.Intelligent drug recommendation systems developed using machine learning-based sentiment analysis have the potential to improve healthcare accessibility and clinical decision-making, as shown here.