A SEMANTIC SENTIMENT ANALYSIS APPROACH TO DETERMINE ROMAN URDU SOCIAL MEDIA COMMENTS

Abstract

This paper presents the second phase of our research study that focused on detection of slang and roman Urdu text in Facebook comment posts. Building upon our previous work, which involved the creation of a comprehensive dictionary and a detailed methodology for word selection and categorization, this study focuses on the actual scientific method of text identification within the comments. We describe the design and implementation of the identification framework and discuss the results comprehensively. A pilot study was conducted to evaluate the efficacy of our approach on a curated dataset of Facebook comment dataset collected over a period of time. The findings highlight both the challenges and potential of Romanized Urdu language and slang detection, offering insights into the linguistic behavior and evolving usage patterns of slang in the digital domain. This study serves as a critical step toward enabling robust content moderation, sentiment analysis, and linguistic preservation especially for low-resource languages.