BSMSD — Bangla Stock Market Sentiment Dataset

Published: 5 August 2026| Version 1 | DOI: 10.17632/crrg2gn2xc.1
Contributors:
Md Mijanur Rahman,
,
,
,

Description

The Bangla Stock Market Sentiment Dataset (BSMSD) is a text-based, manually annotated dataset intended for opinion mining and financial sentiment analysis in Bangla. A main data file with 5,283 distinct text occurrences gathered from online stock market debates makes up the dataset. Dataset description The dataset contains two primary attributes for each instance: • Comment: The original text gathered from social media, discussion boards, or financial comments written in the Bangla script. o Data – Text, Label o Platform – YouTube, Facebook o Stock Name – Full ticker name for which the comment was scraped. • Sentiment Label: The annotated target class represented as a numerical value to its respective sentiment category: o 0 = Neutral , If the text indicate neither negative nor positive sentiment. o 1 = Negative , If the text indicate negative sentiment. o 2 = Positive , If the text indicate positive sentiment.

Files

Institutions

Categories

Computer Science, Economics, Finance, Artificial Intelligence, Computational Linguistics, Data Science, Machine Learning, Business Management

Licence