ASTD: Arabic Sentiment Tweets Dataset
TL
TL;DR
ASTD, an Arabic social sentiment analysis dataset gathered from Twitter, consists of about 10,000 tweets which are classified as objective, subjective positive, subjective negative, and subjective mixed.
Abstract
This paper introduces ASTD, an Arabic social sentiment analysis dataset gathered from Twitter. It consists of about 10,000 tweets which are classified as objective, subjective positive, subjective negative, and subjective mixed. We present the properties and the statistics of the dataset, and run experiments using standard partitioning of the dataset. Our experiments provide benchmark results for 4 way sentiment classification on the dataset.
References8 items
1
Sentiment Analysis For Modern Standard Arabic And Colloquial
2
Building Large Arabic Multi-domain Resources for Sentiment Analysis
3
An Arabic Twitter Corpus for Subjectivity and Sentiment Analysis
4
SANA: A Large Scale Multi-Genre, Multi-Dialect Lexicon for Arabic Subjectivity and Sentiment Analysis
5