Twitter Sentiment Analysis in the Crisis Between Russia and Ukraine Using the Bert and LSTM Model

Sentiment analysis is a method for determining the attitude or viewpoint of individuals towards any trending topic. The most appropriate platform for sharing opinions on any topic is social media. Twitter is a social networking site where users may express their opinions using text messages called tweets. Originally, tweets could only be 140 characters long, but they are now up to 280 characters. Sentiment analysis is a component of NLP that aids in determining the polarity of a text or tweet. Using various keywords like “Ukraine war”, “Ukraine and Russia war”, “Ukraine”, and “NATO in Ukraine,” we are gathering tweets about the conflict between Ukraine and Russia. There were 121156 tweets collected in total, however only 31584 were left for sentiment analysis after cleaning, lemmatization stemming, and deleting duplicate tweets. The “Text Blob” package of Python was used to determine the polarity of the tweets. The present study used the BERT, LSTM, and LSTM with glove methods of machine learning, which produced the best results in comparison to earlier studies with accuracy levels of 97 to 99%.