ParsHate dataset enables Persian hate speech and target detection research

ParsHate: A Benchmark Dataset for Hate and Target Detection in Persian

Computation and LanguageDatabases

Summary

Detecting hate speech helps keep social media safe, but there isn’t enough data in Persian. The authors created ParsHate, a large dataset with 10,000 Persian tweets over ten years, carefully labeled for hateful content and who the hate is directed at. They also marked whether the hate and its targets were explicit or hidden and explained why. Tests with current hate speech detection methods showed there’s room to improve, especially for identifying targets. ParsHate is freely available to help build better tools.

What this means in practice

Authors

Zahra Bokaei, Walid Magdy, Bonnie Webber

Abstract

We introduce ParsHate, a manually annotated dataset of 10,000 Persian tweets spanning 2013-2022, representing the first decade-long benchmark for hate speech detection in Persian. The dataset contains 31% hateful content and supports both hate detection and multi-label fine-grained target identification across seven structured target categories. ParsHate also distinguishes explicit and implicit hate, marks explicit and implicit targets, and provides span-level rationales. Data collection combines random and score-stratified temporal sampling to reduce keyword-driven bias while preserving natural label distributions. Applying SOTA models for Persian hate-speech detection on ParsHate shows moderate performance (79% F1), especially with samples from earlier years, and low performance with target identification (25.5% macro-F1). This emphasizes the diverse sampling of hate speech in ParsHate and its challenging nature that requires more advanced methods for better performance. Dataset is made publicly available.