MaCmS

MaCmS是由高威大学数据科学研究所创建的，首个针对Magahi-Hindi-English代码混合语言的情感分析数据集。该数据集包含5000条来自YouTube的评论，通过手动标注情感极性，旨在解决低资源语言在情感分析中的挑战。数据集的创建过程涉及从YouTube收集数据，去除重复评论，并进行语言特定跨度的标记。MaCmS的应用领域包括跨国公司的客户反馈分析、政治领域的公众情绪评估以及社交媒体平台的冒犯性内容监控，为理解和分析代码混合语言中的情感提供了重要资源。

MaCmS was created by the Data Science Institute at National University of Ireland, Galway. It is the first sentiment analysis dataset dedicated to Magahi-Hindi-English code-mixed languages. This dataset comprises 5,000 YouTube comments that are manually annotated with sentiment polarities, and is developed to tackle the challenges of sentiment analysis for low-resource languages. The construction process of MaCmS includes collecting data from YouTube, eliminating duplicate comments, and conducting language-specific span annotation. Application scenarios of MaCmS cover customer feedback analysis for multinational corporations, public sentiment evaluation in political domains, and offensive content monitoring on social media platforms, thus providing a valuable resource for understanding and analyzing sentiment in code-mixed languages.