Anthonyogunsakin/yoruba-ner-project

# Yoruba Named Entity Recognition (NER) This project explores the development and evaluation of Named Entity Recognition (NER) models on Yoruba news articles. The aim is to improve multilingual NLP and support underrepresented languages like Yoruba. ## Models Used - Logistic Regression - Support Vector Machine (SVM) - Long Short-Term Memory (LSTM) ## Dataset Sourced from the Hugging Face repository. Includes 12,714 entries split into training (80%) and testing (20%). ## Performance Summary - SVM: Highest accuracy (77%) but limited balance across entity types. - Logistic Regression & LSTM: Macro F1-score of 0.53, better across multiple categories. - Person (PER) entities were identified most effectively by all models. ## Contributions This study highlights the limitations of current NER tools for low-resource languages and offers practical improvements. ## File Structure - `notebooks/` - Jupyter notebooks used for training and evaluation. - `models/` - Saved trained models. - `results/` - Performance metrics and visualizations. - `abstract.txt` - Project abstract.