This presentation, delivered at the Quantitative Study on Russian Language seminar in Helsinki in August 2015, detailed a research project that employed semantic vector models to examine Russian poetic language. The focus was on identifying distinct linguistic features within a corpus of Russian poetry.

Objective:

The primary objective was to conduct a comparative linguistic analysis between a specifically curated corpus of Russian poetic texts and a standard Russian language model, as represented by the RusVectores2.0 project.

Methodology:

  • Analysis was based on a corpus comprising 10,000 Russian poetic texts from the period 1890-1920.
  • A vector model, constructed via the word2vec tool, was exclusively developed for this poetic corpus.
  • Comparative analysis with the RusVectores2.0 general Russian language model was undertaken to isolate and identify distinctive poetic linguistic patterns.

Findings:

The comparative analysis elucidated unique semantic attributes inherent to Russian poetic texts. These attributes were markedly distinct from those found in the broader, general Russian language model.

Conclusion:

This study lays the groundwork for subsequent research in digital linguistics and poetic analysis. It underscores the efficacy of semantic vector models in examining the intricate differences between poetic and standard language usage, opening potential pathways for future linguistic inquiry.

To view the complete presentation, please follow this link.