ASAD: A Twitter-based Benchmark Arabic Sentiment Analysis Dataset

TL;DR

This paper introduces ASAD, a large, high-quality Twitter dataset with Arabic sentiment labels, designed for benchmarking sentiment analysis models and supporting a related competition with monetary prizes.

Contribution

It provides a detailed description of a new Arabic sentiment analysis dataset with 95K tweets, including data collection, annotation, and baseline model results.

Findings

ASAD contains 95,000 annotated tweets.

Baseline models establish reference performance levels.

The dataset supports competitive benchmarking in Arabic sentiment analysis.

Abstract

This paper provides a detailed description of a new Twitter-based benchmark dataset for Arabic Sentiment Analysis (ASAD), which is launched in a competition3, sponsored by KAUST for awarding 10000 USD, 5000 USD and 2000 USD to the first, second and third place winners, respectively. Compared to other publicly released Arabic datasets, ASAD is a large, high-quality annotated dataset(including 95K tweets), with three-class sentiment labels (positive, negative and neutral). We presents the details of the data collection process and annotation process. In addition, we implement several baseline models for the competition task and report the results as a reference for the participants to the competition.