jbeno/sentiment_merged

这是一个用于三向情感分类（负面、中性、正面）的数据集，由Stanford Sentiment Treebank (SST-3)和DynaSent Rounds 1和2合并而成。数据集包含102,097个训练样本、5,421个验证样本和6,530个测试样本。标签分布显示训练集中中性样本最多，其次是负面和正面样本。数据集的创建旨在为情感分类模型提供多样化的挑战性评论数据。数据集的语言为英语，采用MIT许可证。

This is a dataset for 3-way sentiment classification of reviews (negative, neutral, positive). It is a merge of Stanford Sentiment Treebank (SST-3) and DynaSent Rounds 1 and 2. The dataset contains 102,097 training examples, 5,421 validation examples, and 6,530 test examples. The label distribution shows that the training set has the most neutral examples, followed by negative and positive examples. The dataset was created to provide diverse and challenging review data for sentiment classification models. The language of the dataset is English, and it is licensed under the MIT license.