аннотация10 (Аннотации)

Koltsova O. Yu., Alexeeva S. V., Kolcov S. N.:An Opinion Word Lexicon and a Training Data set for Russian SentimentAnalysis of Social Media.Automatic assessment of sentiment in large text corpora is an important goalin social sciences. This paper describes a methodology and the results of systemdevelopment for Russian language sentiment analysis. It includes: a publiclyavailable sentiment lexicon, a publicly available test collection with sentimentmarkup and a crowdsourcing website for such markup. The lexicon is aimed atdetecting sentiment in user-generated content (blogs, social media) related to socialand political issues.

Its prototype was formed based on other dictionaries and on thetopic modeling performed on a large collection of blog posts. Topic modelingrevealed relevant (social and political) topics and as a result—relevant words for thelexicon prototype and relevant texts for the training collection. Each word wasassessed by at least three volunteers in the context of three different texts where theword occurred while the texts received their sentiment scores from the samevolunteers as well. Both texts and words were scored from −2 (negative) to +2(positive). Of 7,546 candidate words, 2,753 got non-neutral sentiment scores.The quality of the lexicon was assessed with SentiStrength software bycomparing human text scores with the scores obtained automatically based on thecreated lexicon. 93% of texts were classified correctly at the error level of ±1 class,which closely matches the result of SentiStrength initial application to the Englishlanguage tweets.

Negative classes were much larger and better predicted..