This is a project to train an AI model to learn african names which will help in solving the biases problem from major machine leanring algorithms
# Africa_Name_NER
This is a project to train an AI model to learn african names which will help in solving the biases problem from major machine leanring algorithms
# African Name Recognition (AF-NER)
This project aims to build a Named Entity Recognition (NER) model from scratch, specifically trained to identify African names across diverse languages and regions. It is designed to improve representation and performance of NLP tools for African contexts.
## 🌍 Project Goal
The goal is to:
- Build a deep learning-based NER model capable of recognizing African names.
- Improve existing NLP tools that often miss African personal names due to bias in training datasets.
- Make this project open-source and community-driven to encourage collaboration and contribution.
## 📦 Dataset
We aim to collect a large dataset of African names by:
- Scraping Wikipedia pages on notable African figures by country.
- Crawling public articles, blogs, and government websites listing common African names.
- (Future Work) Incorporating publicly available datasets with African cultural context.
**Data Fields:**
- Full Name
- Country / Ethnicity
- Source URL
- Context Sentence (if available)
> ✅ All data collected respects public content usage guidelines and avoids scraping personal or sensitive information.
---
## 🧠 Training Details
The model will be trained from scratch using:
- Deep Learning (Bi-LSTM, CRF, Transformer, BERT variants)
- Libraries: PyTorch, HuggingFace Transformers, spaCy (for comparison), TensorFlow (optional)
- Evaluation Metrics: Precision, Recall, F1-score
Planned Training Phases:
1. Data Collection & Cleaning
2. Tokenization & Annotation
3. Model Training
4. Evaluation & Comparison
5. Deployment / Integration
## 🚀 How to Run Scripts
1. **Clone this repo:**
```bash
git clone
github.com
cd af-ner
pip install -r requirements.txt
python scripts/scrape_wikipedia.py
python model/train.py
Contributing
We welcome all contributi …