A Sentimental Analysis Approach for Personalized Drug Recommendations Using Machine Learning

Since the virus that causes corona disease was discovered, it has gotten more difficult to find medical professionals with the appropriate licenses, including doctors, nurses, diagnostic equipment, and medications. Due to the intense grieving, a substantial number of people in the medical industry pass away. Because of the scarcity, a lot of individuals began self-medicating without first consulting a doctor, which made the health problem worse. There are numerous well-established applications for machine learning, and recent years have witnessed a growth in the pace and breadth of computerization-related research and development. The goal of this research is to develop a method for recommending medications that might greatly reduce the obligations of specialists. In this work, we built a drug recommendation system that makes estimations about the emotional of patient reviews using a variety of different vectorization techniques, such as using a Bow, The TF-ID Word 2 Vector, or Manually Feature Analysis. The predicted emotions were graded using accuracy, remembrance, flscore, exactness, and the area beneath the slope (AUC). The results show that, when weighed against the other models, the classification technique Linear SVC + The TF -I vector process has the greatest accuracy.