Attitudes Toward Feminism in Ukraine: A Sentiment Analysis of Tweets

Abstract

This paper presents a sentiment analysis of Ukrainian tweets on feminism. In order to carry out a computational study of opinions, we have adjusted the SentiStrength algorithm to the Ukrainian language by replacing the English term lists in the program files with the Ukrainian ones. The main contribution is an attempt to compile a social domain sentiment lexicon for Ukrainian (3,736 words). The SentiStength output has shown a prevailing negative sentiment of the analyzed tweets. The program performance was evaluated in terms of accuracy, precision, recall, error, fallout and F1 Score. In addition, we found a number of common attributes of a feminist, which also predominantly express negative attitude. Overall, the findings show that a direct support of a key feminist goal, i.e. equality of women and men in society, by the Ukrainian Tweeter users couples with misconception about the concept of feminism and unwillingness to be called a feminist.