Unified Bangla Multi-class Emotion Corpus (UBMEC)

UBMEC是由山东财经大学和深圳大学联合创建的多类别情感分析数据集，旨在解决孟加拉语情感分类问题。该数据集整合了两个已有的孟加拉语情感数据集BNEmo和BEmoC，并新增了由研究团队手动标注的评论，总计包含13,072条评论，涉及愤怒、厌恶、恐惧、喜悦、悲伤和惊讶六种情感类别。数据集来源于多个领域，如食品、软件、娱乐等，通过精细的预处理和标注过程构建而成，适用于机器学习和深度学习模型的训练与评估，特别是在资源受限的语言环境中。

UBMEC is a multi-class sentiment analysis dataset jointly created by Shandong University of Finance and Economics and Shenzhen University, aiming to address Bengali sentiment classification tasks. This dataset integrates two existing Bengali sentiment datasets, BNEmo and BEmoC, and adds manually annotated comments from the research team, with a total of 13,072 comments covering six sentiment categories: anger, disgust, fear, joy, sadness, and surprise. The dataset is sourced from multiple domains such as food, software, entertainment and other fields, and is constructed through rigorous preprocessing and annotation processes, suitable for training and evaluating machine learning and deep learning models, especially in low-resource language environments.