Drug Recommendation System based on Sentiment Analysis of Drug Reviews using Machine Learning

Authors

  • Mrs. A N RAMAMANI Author
  • PECHETTI JAI KRISHNA Author

Keywords:

evaluated by precision, recall, f1score, accuracy, and AUC score

Abstract

Since coronavirus has shown up, 
inaccessibility of legitimate clinical 
resources is at its peak, like the 
shortage of specialists and 
healthcare workers, lack of proper
equipment and medicines etc. The 
entire medical fraternity is in distress, 
which results in numerous 
individual’s demise. Due to 
unavailability, individuals started 
taking medication independently 
without appropriate consultation, 
making the health condition worse 
than usual. As of late, machine 
learning has been valuable in 
numerous applications, and there is 
an increase in innovative work for 
automation. This paper intends to
present a drug recommender system 
that can drastically reduce specialists 
heap. In this research, we build a 
medicine recommendation system 
that uses patient reviews to predict 
the sentiment using various 
vectorization processes like Bow, TFIDF,Word2Vec, and Manual Feature 
Analysis, which can help recommend 
the top drug for a given disease by 
different classification algorithms.
The predicted sentiments were 
evaluated by precision, recall, 
f1score, accuracy, and AUC score. 
The results show that classifier 
LinearSVC using TF-IDF vectorization 
outperforms all other models with 
93% accuracy.

Downloads

Published

08-06-2024

How to Cite

Drug Recommendation System based on Sentiment Analysis of Drug Reviews using Machine Learning. (2024). International Journal of Engineering Research and Science & Technology, 20(2), 1086-1100. https://ijerst.org/index.php/ijerst/article/view/374