Semantic Rule-Based Sentiment Detection Algorithm for Russian Publicism Sentences

Abstract

This article studies the task of sentiment detection in Russian sentences, which is understood as the author’s attitude on the sentence topic expressed through linguistic expression features. Today most studies on this subject utilize texts of a colloquial style, limiting the applicability of their results to other styles of speech, particularly to publicism. To fill the gap, the authors developed new publicism sentences oriented toward a sentiment detection algorithm. The algorithm recursively applies appropriate rules to parts of sentences represented as constituency trees. Most of the rules are proposed by a philologist, based on knowledge of expression features from Russian philology, and are algorithmized using constituency trees generated by the algorithm. A decision tree and sentiment vocabulary are also used in this study. This article contains the results of evaluation of the algorithm on the corpus of publicism sentences OpenSentimentCorpus and the F-measure is 0.80. The results of errors analysis are also presented.