Evaluating Cross-Lingual Classification Approaches Enabling Topic Discovery for Multilingual Social Media Data

TL;DR

This paper compares various cross-lingual classification methods to improve topic discovery in multilingual social media data, focusing on hydrogen energy discussions across four languages over a decade.

Contribution

It systematically evaluates four cross-lingual approaches for filtering relevant social media content and extracting topics, providing insights into their trade-offs and effectiveness.

Findings

Translation-based models improve relevance filtering in some languages.

Multilingual transformers offer direct cross-lingual classification without translation.

Hybrid strategies balance translation and multilingual approaches for optimal results.

Abstract

Analysing multilingual social media discourse remains a major challenge in natural language processing, particularly when large-scale public debates span across diverse languages. This study investigates how different approaches for cross-lingual text classification can support reliable analysis of global conversations. Using hydrogen energy as a case study, we analyse a decade-long dataset of over nine million tweets in English, Japanese, Hindi, and Korean (2013--2022) for topic discovery. The online keyword-driven data collection results in a significant amount of irrelevant content. We explore four approaches to filter relevant content: (1) translating English annotated data into target languages for building language-specific models for each target language, (2) translating unlabelled data appearing from all languages into English for creating a single model based on English annotations, (3) applying English fine-tuned multilingual transformers directly to each target language data, and (4) a hybrid strategy that combines translated annotations with multilingual training. Each approach is evaluated for its ability to filter hydrogen-related tweets from noisy keyword-based collections. Subsequently, topic modeling is performed to extract dominant themes within the relevant subsets. The results highlight key trade-offs between translation and multilingual approaches, offering actionable insights into optimising cross-lingual pipelines for large-scale social media analysis.