Mapping Ethnic Attitudes in the Russian blogosphere
Project Participants: S. Nikolenko and A. Shimorina
The project ‘Mapping Ethnic Attitudes in the Russian blogosphere’ is dedicated to mapping the attitudes of the Russian-language bloggers (Russian-language Livejournal segment taken as the case) to various ethnic communities, both within and outside Russia. The project is based on mixed-method approach and is also aimed at testing new text clusterisation methods. In particular, topic modeling based on LDA algorithm is tested as the clusterisation method for analysis of blog posts. In terms of methodology, the project aims two major goals. First, we’re working upon substantial enhancing of the LDA algorithm in terms of automatic location of discussion topics. Today, even the definition of an automatically defined ‘topic’ is blurred; other aspects of topic modeling, such as stability of topic recognition or interrelation of pre-defined number of topics with dataset characteristics, are also understudied. To move the modeling nearer to ‘natural’ understanding of discussion topics, additions to the LDA algorithm are proposed: topic stability tests and a new ISLDA method of topic extraction need to be tested. The second methodological goal is to create and test a method for semi-automated cognitive frame analysis based on manual coding of n-grams within a proposed polarization frame. In terms of substance, we aim at mapping the attitudes within a polarization frame that interprets perceived ethnicities as either ‘political’ or ‘cultural’ (‘political actors’ or ‘ritualized Others’) and to suggest the causes of such a polarization in public mind.
1. Bodrunova S. Working materials on ethnic analysis of blogs. LINIS PowerPoint presentation, September 2013.
2. Mapping Ethnic Attitudes in the Russian Blogosphere: ISLDA as a Method of Social Research on the Internet [anonymous submission; S.Bodrunova, S.Nikolenko, S.Koltsov, O.Koltsova, A.Shimorina] / ICA Annual Meeting paper. 2013.
S. Bodrunova, S. Koltsov, O. Koltsova, S.I. Nikolenko, A. Shimorina. Interval Semi-Supervised LDA: Classifying Needles in a Haystack. Proceedings of the 12th Mexican International Conference on Artificial Intelligence (MICAI 2013) // Lecture Notes in Artificial Intelligence. Vol. 8265. P. 265-274. [Web of Science, Scopus].
ISLDA – software for interval semi-supervised LDA algorithm-based topic modeling plus its web interface (under construction). Developer: S.I.Nikolenko.
Have you spotted a typo?
Highlight it, click Ctrl+Enter and send us a message. Thank you for your help!
To be used only for spelling or punctuation mistakes.