Related Experiment Video
Updated: Feb 4, 2026

Integrating Computerized Linguistic and Social Network Analyses to Capture Addiction Recovery Capital in an Online Community
Published on: May 31, 2019
Social media discourse on feminism: A dataset for sentiment analysis in bangla comments
Md Mijanur Rahman1, Md Sumon Hosen1, Zaid Bin Sajid1
1Department of Computer Science and Engineering, Southeast University, 251/A Tejgaon I/A, Dhaka 1208, Bangladesh.
Abstract:
Bangladesh is a socio-culturally diverse country where perspectives on women's freedom vary significantly. Social media sites are now important places for sharing feminist ideas through public comments and posts. This study offers a detailed collection of 6,830 comments in Bangla about feminism to help analyze public opinion. Most of this data comes from Facebook, with some also from Instagram and Twitter. Data collection involved systematic extraction from public groups and targeted hashtag searches, including (women's rights). Native Bangla speakers meticulously annotated each comment by hand to ensure that it was topical and to identify any abusive language. This manual validation procedure guarantees a high-quality dataset appropriate for the study of online gender-based violence in the Bangla language context, sentiment analysis, and abusive language analysis, among other machine learning and NLP tasks. In addition, comments were divided into three sentiment classes: neutral, negative, and positive. This allowed for thorough analysis of feminist discourse on Bangladeshi social media and supervised learning.
Related Concept Videos
Social Scripts
Social Proof
Social Traps
Social Exchange Theory
Social Facilitation
Social Loafing

