A Transformer Based Medicine Recommendation System that Uses Drug Reviews

Although the internet is a vast source of information, caution must be exercised to avoid accessing harmful content. The abundance of clinical data scattered across numerous websites can make it difficult for users to find useful information that can improve their health and wellbeing. Studies indicate that nearly 60% of adults search for health information on the internet, with 35% of them solely focusing on diagnosing their ailments online. Medical errors resulting from doctors’ limited experiences in prescribing medication have been linked to several deaths. To address this issue, a medicine recommendation system has been proposed for doctors to use while prescribing medication. Medication errors are a significant risk to patients’ lives and are among the most serious medical errors. This highlights the need for health advisory systems to help users make more accurate and effective decisions about their health. We found that in the special case of healthcare recommendations where the accurate results are a necessity, utmost importance should be given to the data in hand and how that data is used for assistance instead most of the time the focus is shifted to the model and its complexity. We propose a recommendation system that has a less computationally intensive process when compared to similar other ones with the same motive, our system analyzes the data thoroughly than the existing research, find the polarity of the drug reviews, and aims to make accurate recommendations of drugs for the asked condition almost instantly using a novel approach. This kind of recommender system underscores the importance of technology in the modern world, particularly in the medical field where it can potentially save lives by assisting doctors.