Rule-Based Sentiment Analysis of Ukrainian Reviews

Abstract

Last decade witnessed a lot of research in the field of sentiment analysis. Understanding the attitude and the emotions that people express in written text proved to be really important and helpful in sociology, political science, psychology, market research, and, of course, artificial intelligence. This paper demonstrates a rule-based approach to clause-level sentiment analysis of reviews in Ukrainian. The general architecture of the implemented sentiment analysis system is presented, the current stage of research is described and further work is explained. The main emphasis is made on the design of rules for computing sentiments.