A Comparative Analysis of Convolutional and Recurrent Neural Networks for Multi-Label Bangla News Classification on BARD Corpus

This dataset contains Bangla-language news articles collected from various online news portals and websites, representing five distinct categories of real-world news content. It has been designed to support machine learning and deep learning tasks related to text classification, natural language processing (NLP), and computational linguistics in the Bangla language.
The news articles have been collected from popular and widely-read Bangla news platforms such as Prothom Alo, Jugantor, Ittefaq, Bdnews24, Kaler Kantho, Bangla Tribune, and Samakal. These sources ensure a diverse and representative coverage of current events and topics in Bangladesh.
The primary purpose of this dataset is to facilitate the development and benchmarking of models that can accurately classify Bangla news articles into their respective topical categories. Unlike sentiment or opinion datasets, this collection focuses on categorical news classification to assist in automating news organization, topic filtering, and personalized content delivery systems. The dataset can also be used to support linguistic research and Bangla NLP pipeline development.
The dataset is organized into five balanced classes, each representing a different news domain:
Economy: 500 articles
Entertainment: 500 articles
International: 500 articles
Sports: 500 articles
State: 500 articles
Total Number of Articles: 2,500
Language: Bangla
File Format: Plain text (.txt)