83 datasets found

The Toxic Comment Classification Challenge dataset contains comments from Wikipedia organized in six classes: toxic, severe toxic, obscene, threat, insult, and identity hate.

The Hate Speech Tweets dataset contains over 24,000 English tweets labeled as non-offensive, hate speech, and profanity.

A dataset of 150 research papers from the domain of computer science, manually annotated and class labelled for sentiment analysis.

Stream TwitterSentiment is a dataset of tweets, focusing on sentiment analysis, and is used to test the performance of active stream learning algorithms for polarity learning.

StreamJi is a dataset of product reviews, focusing on features of products, and is used to test the performance of active stream learning algorithms for polarity learning.

The Hatespeech dataset is a collection of tweets containing lexicons used in hate speech.

IMDB and Yelp datasets for sentiment classification

IMDB and Yelp datasets for sentiment classification

IMDB and Yelp are datasets used for sentiment analysis and author identification.

The dataset is used for entity-specific sentiment classification of Yahoo News comments.

Sentiment training dataset for LABDet, a robust and language-agnostic bias probing method to quantify intrinsic bias in monolingual PLMs.

The Tweet Sentiment Extraction dataset contains positive, negative, and neutral tweets with human-annotated rationales.

The Movie Reviews dataset contains positive and negative movie reviews with rationales annotated by humans to support classification.

The DynaSent dataset contains approximately 122,000 sentences, each labeled as positive, neutral, or negative.

MoodyLyricsPN is a bigger collection of 5000 songs labeled as positive or negative only.

MoodyLyrics4Q is a dataset of 2,000 songs, fully compliant with the four requisites listed in the previous section.

VideoIC dataset for automatic live video commenting

Livebot dataset for automatic live video commenting

Sentiment-oriented Transformer-based Variational Autoencoder (So-TVAE) for Live Video Commenting

The dataset used in the paper is the New York Times dataset and the 20Newsgroups dataset.