Gibberish Semantics: How Good is Russian Twitter in Word Semantic Similarity Task?

TL;DR

This paper evaluates the effectiveness of Russian Twitter data for word semantic similarity tasks, comparing it to other corpora and highlighting its potential as a valuable resource for language modeling.

Contribution

It introduces Twitter as a new corpus for Russian semantic similarity tasks and compares its performance with other existing corpora.

Findings

Twitter-based vectors perform comparably to other single-corpus models

Combining multiple corpora yields the best semantic similarity results

Twitter corpus is a viable resource for Russian language modeling

Abstract

The most studied and most successful language models were developed and evaluated mainly for English and other close European languages, such as French, German, etc. It is important to study applicability of these models to other languages. The use of vector space models for Russian was recently studied for multiple corpora, such as Wikipedia, RuWac, lib.ru. These models were evaluated against word semantic similarity task. For our knowledge Twitter was not considered as a corpus for this task, with this work we fill the gap. Results for vectors trained on Twitter corpus are comparable in accuracy with other single-corpus trained models, although the best performance is currently achieved by combination of multiple corpora.