Twitter Based Sentiment Analysis of Russia-Ukraine War Using Machine Learning

Abstract

Social media platforms and micro blogging websites can be used as a potential source for gathering opinions and sentiments from the public on a variety of topics, such as the present state of affairs in nations that have experienced conflict. Twitter, in example, offers a variety of text tweets that might link to feelings across time and geography. Using Textblob and Vader as a lexicon method, this research paper performs sentiment analysis over a dataset containing tweets regarding the situation before and after Russia invades Ukraine. It also performs standard machine learning over the dataset. This machine learning model categorizes opinions about Russia's invasion of Ukraine according to sentiments. The current study examines different machine learning algorithms and focuses on the Doc2Vec feature extraction approach utilizing Chi2 (χ2) as a feature selection. The objective of this research is to use Twitter to get people's opinions about the war. The current study helps news media organizations analyze public opinion, particularly that of Russia and Ukraine, about the conflict and draw attention to upcoming difficulties.