Potrika

Potrika是由阿卜杜勒阿齐兹国王大学高性能计算中心创建的大型单标签孟加拉语新闻文章数据集，旨在支持自然语言处理研究。该数据集从孟加拉国的六个流行在线新闻门户收集，涵盖2014至2020年的新闻文章，包含185.51百万字和12.57百万句子，分为八个类别，每个新闻文章附带五个属性。Potrika不仅包括原始数据集，还通过NLP增强技术创建了平衡数据集，适用于文本分类、摘要生成等多种NLP研究。数据集的应用领域广泛，旨在解决孟加拉语新闻分类等问题的研究需求。

Potrika is a large-scale single-label Bengali news article dataset developed by the High Performance Computing Center at King Abdulaziz University, intended to support natural language processing (NLP) research. Collected from six popular online news portals in Bangladesh, the dataset spans news articles from 2014 to 2020, totaling 185.51 million words and 12.57 million sentences, and is categorized into eight classes, with each news article associated with five attributes. Potrika includes both the original dataset and a balanced dataset constructed via NLP augmentation techniques, which is applicable to a variety of NLP research tasks such as text classification and text summarization. The dataset has a wide range of application scenarios, aiming to address the research needs of tasks such as Bengali news classification.