The objective of this project is to carry out sentimental analysis on commonly prescribed anti-hypertensive drugs in the united state. Build a web-app that reassure user of these drugs based on review sentiment. Anti hypertensive drugs used are:
- Amlodipine
Brand names: Norvasc, Katerzia, Norliqva - Metoprolol
Brand names: Metoprolol Succinate ER, Metoprolol Tartrate, Lopressor, Toprol-XL - Lisinopril
Brand names: Zestril, Prinivil, Qbrelis - Losartan
Brand names: Cozaar - Furosemide
Brand names: Lasix, Diaqua-2, Lo-Aqua
- The data set used in notebook was scraped from
- Scraping script can be found Here.
Challenges faced while getting these review sentitments are:
- Most user are using these drug with combination of other drugs for other co-morbidities.
- Users are on diffrent dosage.
- Druration of use are diffrent for users.
- Some review are sarcastic and aspected base.
- Most expected side effect are consider as adverse effect by users.
Over all, Textbolb sentiment polarity and Nltk sentiment analyser did perform better than Bert.
