smaliyu/NaijaOffens

This is a repositorty for the NaijaHate. A Data Science Africa (DSA) funded project for the creation of Hate speech corpus for the 3 main Nigerian languages, Hausa, Yoruba and Igbo. # NaijaHate: Offensive Language and Hate Speech Detection in Nigerian Languages ## Introduction NaijaHate is a pioneering project aimed at addressing the growing problem of online offensive and hate speech in Nigeria. This project introduces the first labeled dataset for offensive and hate speech detection in three major Nigerian languages. Our goal is to aid in content moderation, social media monitoring, and threat evaluation through machine learning tools. ## Table of Contents - Project Description - Dataset - Lexicon - Model Training - Research Paper - Project Sponsors - Contact ## Project Description The Offensive and Hate Speech Detection Project aims to tackle the pervasive issue of offensive and hateful content circulating in Nigeria's digital spaces. Leveraging state-of-the-art Natural Language Processing (NLP) techniques, the project seeks to develop robust algorithms and tools capable of automatically identifying and mitigating instances of hate speech across the three major languages spoken in Nigeria. ### Challenges Faced in Nigeria and the Role of Hate Speech Detection Projects 1. **Cultural Diversity**: Nigeria boasts over 250 ethnic groups, each with distinct languages and customs. This diversity can sometimes lead to tensions exacerbated by offensive or hateful speech. Offensive and Hate speech detection projects can help identify and mitigate such instances, fostering understanding among Nigeria's diverse population. 2. **Linguistic Variation**: With over 500 languages spoken, linguistic diversity poses a challenge for automated text analysis. Hate speech detection models must accurately identify offensive language across multiple languages and dialects, necessitating development in various Nigerian languages. 3. **Legal and Regulatory Framework**: Nigeria has laws against hate speech, but enforcement can be difficult, especially online. Hate speech detection projects can complement legal frameworks by aiding in the identification and report …