A Drug Recommender System based on Sentiment Analysis of Drug Reviews using Deep Learning

Abstract

In today’s digital world, the younger to older generation is dependent on internet not only for entertainment, shopping but also to live a healthy life. Artificial Intelligence and Machine Learning has also made its place in healthcare. The health forum data and the data extracted from social mediaare major sources of the experiences shared on chronic diseases and drugs.A major percentage of users are searching on theinternetfor data related to wellbeing. But study shows that most of the people die by taking wrong medicine. Tons of informationis there on theinternet which can be helpful but can also be hazardous. There are researchers who have dealt with areas like electronic items, films and restaurants yet very few have worked widely on healthcare and medical domains.In this paper, Natural Language Processing techniques are applied on the patient reviews to classify the emotions and recommend the top medicines to the patient. In this paper, CNN-LSTM gives the highest accuracy (83.06 %) than any other ML & DL Model. This paper presents that DL model works better than ML models on this dataset.