Sentiment analysis on twitter tweets about COVID-19 vaccines usi ng NLP and supervised KNN classification algorithm

F. M. Javed Mehedi Shamrat (Daffodil International University)
Sovon Chakraborty (European University of Bangladesh)
M. M. Imran (Cefalo Bangladesh Limite)
Jannatun Naeem Muna (European University of Bangladesh)
Md. Masum Billah (Daffodil International University)
Protiva Das (BRAC University)
Md. Obaidur Rahman (European University of Bangladesh)

Indonesian Journal of Electrical Engineering and Computer Science
Vol 23, No 1: July 2021

The pandemic has taken the world by storm. Almost the entire world went into lockdown to save the people from the deadly COVID-19. Scientists around the around have come up with several vaccines for the virus. Amongthem, Pfizer, Moderna, and AstraZeneca have become quite famous. General people however have been expressing their feelings about the safety and effectiveness of the vaccines on social media like Twitter. In this study, such tweets are being extracted from Twitter using a Twitter API authentication token. The raw tweets are stored and processed using NLP. The processed data is then classified using a supervised KNN classification algorithm. The algorithm classifies the data into three classes, positive, negative, and neutral. These classes refer to the sentiment of the general people whose Tweets are extracted for analysis. From the analysis it is seen that Pfizer shows 47.29%positive, 37.5% negative and 15.21% neutral, Moderna shows 46.16%positive, 40.71% negative, and 13.13% neutral, AstraZeneca shows 40.08%positive, 40.06% negative and 13.86% neutral sentiment.