mnemoraorg/tweetfeels-1m6

TweetFeels 1m6是一个包含超过一百万条英语推文的集合，这些推文通过Twitter API在2009年收集。每条推文根据表情符号自动推断出情感标签（0代表负面，2代表中性，4代表正面）。数据集以CSV文件形式提供，包含推文情感极性、ID、发布日期、搜索查询、用户名和文本字段。这个数据集没有经过人工标注，标签是基于表情符号的存在通过远监督方法分配的，适合用作情感分析和社交媒体文本挖掘的大规模基准。

TweetFeels 1m6 is a collection of over a million English-language tweets harvested through the Twitter API in 2009. Each tweet is assigned a sentiment label automatically inferred from the presence of emoticons (0 for negative, 2 for neutral, and 4 for positive). The dataset is provided as a CSV file containing fields for sentiment polarity, tweet ID, posting date, search query, user screen name, and full tweet text. The labels are assigned through distant supervision based on the presence of emoticons without manual annotation, making it a standard benchmark for large-scale sentiment analysis and social-media text-mining tasks.