Open database and a crowdsourcing website for development of sentiment analysis instruments (2014-2015)

Project Head: Sergei Koltcov
Project Participants: Svetlana Alexeyeva, Olessia Koltsova

This project is supported by the Russian Humanitarian Scientific Fund in 2014. The project is aimed at the development of a software, a database and a sentiment lexicon that can further help create a methodology of detecting emotional polarity of user-generated Russian language content. Combined with automatic topic detection it will allow social scientists to mine public opinion from user texts  - or, more precisely, attitudes of the Internet-active fraction of population to socially important issues.  The project has the following goals:

1. To retrieve potentially emotional words from large databases of blog content collected in the Laboratory.
2. To make sentiment mark-up of this draft lexicon by volonteers at a crowdsourcing website.
3. To develop such a crowsourcing website that later will be available to researchers for crowdsourcing mark-up of other lexicons and thesauri;
4. To develop a benchmark text collection and to test the quality of thelexicon.

As a result the tested sentiment lexicon will be available at the developed crowdsourcing website together with possibilities to carry out new research.  This will solve a fundamental problem of absence of the Russian-language sentiment lexicon for user-generated texts in the domain of public affairs.