Drug Recommendation System using Drug Reviews and Useful Count

In this era, we have recently seen the very dangerous spread of viruses all over the world called coronavirus. During this period we have seen there was a lack of hospital beds, doctors, healthcare workers, and lack of equipment. So many people died every day and night. The death rate was increasing like hell. Due to this situation, many people started to take medicine without any doctor consultation, but people did not know which medicine and how much dosage has to be consumed. That made the situation worsens. As we know computer technologies are growing rapidly, why can’t we make use of these technologies to eradicate the worst situation like coronavirus spread? Yes, we can go for Machine Learning algorithms [3] which are having numerous applications. And so many automation works are running on machine learning algorithms [3]. This work is to develop a drug recommendation system [8] based on machine learning algorithms [3], that can help to reduce the specialist heap. In this work, we build the drug recommendation system by using the patient’s reviews and useful counts of reviews to predict the sentiment using the various vectorization [2] methods such as Bow, TF-IDF, Word2Vec [2]. We build and compare the classification models such as ANN [8] and LSTM [6]. The best result is considered and used for the Drug Recommendation System. We recommend the top drug for a given disease. Finally, our work is going to be evaluated by precision, recall, flscore, accuracy score [2]