Rule-Based Translation of Sinhala Slang and Colloquial Expressions into English for Enhanced Cross-Cultural Communication

Abstract

Begin This research paper presents the development and evaluation of a hybrid translation system designed to translate Sinhala slang and colloquial expressions into English. Recognizing the challenges of accurately conveying informal language in cross-cultural communication, the study addresses the need for precise and contextually relevant translations. Sinhala slang and colloquialisms often contain nuanced meanings, cultural references, and idiomatic expressions that are challenging to capture with conventional machine translation approaches. To address these complexities, the proposed system combines linguistic rules, pattern recognition algorithms, and context-based translation models, specifically tailored for Sinhala slang and colloquial styles. A novel component of this research is the combination of rule-based matching (for known slang) with unsupervised learning using Word2Vec embeddings and K-Means clustering for new slang detection. This hybrid approach enhances the system’s ability to differentiate formal from informal language by leveraging predefined rules for familiar slang while dynamically identifying emerging slang patterns through clustering. This integration enables more targeted translation processing by identifying slang, which is subsequently handled by rule-based frameworks developed through data collection and analysis of Sinhala slang expressions. The system’s effectiveness is rigorously evaluated using metrics such as accuracy, precision, recall, F1-score, and Bilingual Evaluation Understudy score, confirming its utility in promoting cross-cultural understanding through culturally sensitive translations. The outcomes of this research advance rule-based and unsupervised learning-supported machine translation techniques for informal language, fostering communication across diverse linguistic and cultural contexts.