short_text

This collection includes multiple short text classification datasets designed for various natural language processing tasks. It contains several topic classification datasets, such as AG'News, Snippets, and TMNNews, which cover a wide range of topics and domains to evaluate the effectiveness of classification models. Additionally, the collection includes a binary sentiment classification dataset, such as Twitter, aimed at determining positive or negative sentiment in text. All datasets are provided in CSV format, with the first column representing the labels (e.g., categories or sentiments) and the second column containing the corresponding text content. These datasets are particularly useful for benchmarking machine learning models and conducting experiments on both topic and sentiment classification tasks. They cover diverse real-world applications, such as news categorization, domain-specific snippet classification, and social media sentiment analysis, making them versatile resources for research and development in computational linguistics.

本数据集集合包含多款专为各类自然语言处理任务设计的短文本分类数据集。其中涵盖多款主题分类数据集，例如AG'News、Snippets与TMNNews，这些数据集覆盖广泛的主题与领域，可用于评估分类模型的性能有效性。此外，该集合还包含一款二元情感分类数据集Twitter，旨在实现文本正负向情感的判断。所有数据集均以CSV格式提供，第一列为标签（例如类别或情感标签），第二列为对应的文本内容。此类数据集尤其适用于机器学习模型的基准测试，以及主题分类与情感分类任务的相关实验。它们覆盖了新闻分类、领域专属片段分类、社交媒体情感分析等多样真实应用场景，是计算语言学领域研究与开发工作中通用性极强的优质资源。