ANALYZING CODE-SWITCHING PATTERNS IN MULTILINGUAL SOCIAL MEDIA CORPORA: A COMPUTATIONAL LINGUISTICS APPROACH

Abstract

Code-switching, the alternation between two or more languages in a single conversation or text, is a prevalent phenomenon in multilingual communities. Despite its importance, code-switching remains underexplored in digital communication, particularly in social media. This study addresses this gap by analyzing code-switching patterns in multilingual social media corpora using computational linguistics techniques. We curate a large-scale, annotated corpus of social media text and develop transformer-based models to identify and classify code-switching points. Our analysis reveals insights into the linguistic and social factors influencing code-switching behavior, including language proficiency, topic, and sentiment. The study sheds light on the complex dynamics of multilingual language use in digital communication, with implications for language technology, sociolinguistics, and multilingual communication studies. The findings contribute to a deeper understanding of code-switching in social media and inform the development of more effective language processing tools.