Year of Graduation
Creation of a Sentiment Lexicon from Domain Social Media for Sentiment Analysis
Language Theory and Computational Linguistics
The work is devoted to creation of a sentiment lexicon for sentiment-analysis of the comments in the online-game community of the social network VK. The work provides a multidisciplinary analysis of the texts, a process of creation of a training sample, methods for identifying sentiment lexicon, the results of applying machine learning for sentiment-analysis and a rule-based approach. The statistical algorithm based on the Random Forest Classifier was improved with the help of rules based on the obtained lexicon.