Twitter Emotion Dataset
收藏arXiv2025-09-30 收录
下载链接：
https://medium.com/age-of-awareness/how-to-use-wheel-of-emotions-to-express-better-emotions-8037255aa661
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了一系列按六种主要情绪分类的推文：愤怒、糟糕、害怕、快乐、悲伤和惊讶。为了通过情感分析消除错误标记的推文，该数据集经过了两个预训练的情感分析网络——BERTSent和XLM-roBERTa-base的筛选，以确保错误标记的推文数量最小。基于此数据集训练的模型在测试阶段的准确率超过了90%，任务主要是情感分析和情绪分类。
This dataset comprises a collection of tweets categorized into six primary emotions: anger, disgust, fear, joy, sadness, and surprise. To eliminate mislabeled tweets via sentiment analysis, this dataset was filtered by two pre-trained sentiment analysis networks—BERTSent and XLM-roBERTa-base—to minimize the number of mislabeled tweets. Models trained on this dataset achieved an accuracy exceeding 90% during the test phase, with the core tasks being sentiment analysis and emotion classification.
提供机构：
Twitter
搜集汇总
数据集介绍
以上内容由遇见数据集搜集并总结生成
