RuSentiment: An Enriched Sentiment Analysis Dataset for Social Media in Russian

Conference · Mon Aug 20 00:00:00 EDT 2018

OSTI ID:1529951

Rogers, Anna ^[1]; Romanov, Alexey ^[1]; Rumshisky, Anna ^[1]; Volkova, Svitlana ^[2]; Gronas, Mikhail ^[3]; Gribov, Alex ^[1]

---

1. University of Massachusetts at Lowell
2. BATTELLE (PACIFIC NW LAB)
3. Dartmouth College

This paper presents RuSentiment, a new dataset for sentiment analysis of social media posts in Russian, and a new set of comprehensive annotation guidelines that are extensible to other languages. RuSentiment is currently the largest in its class for Russian, with 30,521 posts annotated with Cohen’s kappa of 0.58 (3 annotations per post). To diversify the dataset, 6,749 posts were pre-selected with an active learning-style strategy. We report baseline classification results, and release the bestperforming embeddings trained on 3.2B tokens in Russian VKontakte posts.

• Research Organization: Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
• Sponsoring Organization: USDOE
• DOE Contract Number: AC05-76RL01830
• OSTI ID: 1529951
• Report Number(s): PNNL-SA-134041
• Resource Relation: Conference: Proceedings of the 27th International Conference on Computational Linguistics (COLING 2018), August, 2018, Santa Fe, NM
• Country of Publication: United States
• Language: English

Related Subjects

data collection
sentiment analysis
machine learning
natural language processing