Published research / 2020

Arabic hate-speech detection research

A dated research milestone on Arabic short-text classification, human annotation, preprocessing and model comparison.

Published research
View sourceDiscuss your use case

Overview

The research method

  • Input
  • Human
  • Rule / logic
  • AI step
  • Output
  1. 01Input
    Gather

    Data collected September–October 2019

  2. 02Human
    Annotate

    Human annotation

  3. 03Rule / logic
    Preprocess

    Arabic-specific preprocessing

  4. 04AI step
    Compare

    Classical and neural methods

  5. 05Output
    Publish

    Paper published in 2020

As documented in the 2020 paper 'Hate Speech Detection of Arabic Shorttext'. Sensitive example text is not reproduced.
01

Research context

The 2020 paper 'Hate Speech Detection of Arabic Shorttext' names Khaled Taha and Mahmoud Al-Sharif of Trafalgar AI's Social Media Lab. It documents data gathering in September–October 2019, human annotation, Arabic preprocessing and comparison of classical and neural methods.

02

Attribution and limits

This is an attributed Trafalgar AI research milestone. It does not establish that the current Toot entity independently developed every component, nor does a historical paper provide present-day product accuracy figures. Sensitive example text is not reproduced.

Evidence and status

Published research

Published research with attributed authors

Published, attributed research.

Source: www.csitcp.com/paper/10/105csit07.pdf

Talk to us about the problem.

Start with the workflow, your constraints and what a successful implementation should achieve.

Get in touchExplore products