Automated Detection of Non-Relevant Posts on the Russian Imageboard "2ch": Importance of the Choice of Word Representations

TL;DR

This paper investigates how different word embedding models affect the automated detection of non-relevant posts on Russian forums by comparing their performance on a specially created dataset.

Contribution

It introduces a comparative analysis of seven word embedding models for semantic relatedness detection in Russian forum posts, highlighting the importance of embedding choice.

Findings

FastText and Swivel outperform other models in relatedness detection

The dataset reveals challenges due to Russian lexical and grammatical features

Word embedding choice significantly impacts detection accuracy

Abstract

This study considers the problem of automated detection of non-relevant posts on Web forums and discusses the approach of resolving this problem by approximation it with the task of detection of semantic relatedness between the given post and the opening post of the forum discussion thread. The approximated task could be resolved through learning the supervised classifier with a composed word embeddings of two posts. Considering that the success in this task could be quite sensitive to the choice of word representations, we propose a comparison of the performance of different word embedding models. We train 7 models (Word2Vec, Glove, Word2Vec-f, Wang2Vec, AdaGram, FastText, Swivel), evaluate embeddings produced by them on dataset of human judgements and compare their performance on the task of non-relevant posts detection. To make the comparison, we propose a dataset of semantic relatedness with posts from one of the most popular Russian Web forums, imageboard "2ch", which has challenging lexical and grammatical features.