Sultandammy/nigerian-language-identifier

A machine learning application for identifying English, Hausa, Igbo, and Yoruba text. \# 🇳🇬 Nigerian Language Classifier A machine learning application that identifies whether a short text is written in English, Hausa, Igbo, or Yoruba. Built as a capstone project for the \*\*3MTT / NextGen AI \& Machine Learning\*\* program. \## 📌 Project Overview Language identification is an important component of many applications, including chatbots, translation systems, search engines, and other natural language processing (NLP) applications. This project develops a machine learning classifier that can identify four languages commonly used in Nigeria: \- 🇬🇧 English \- 🇳🇬 Hausa \- 🇳🇬 Igbo \- 🇳🇬 Yoruba The application accepts a short text phrase and predicts the most likely language. \## 🎯 Project Objective The main objective is to build a working language classification system that can: \- Accept a short text phrase as input \- Identify the language of the text \- Evaluate model performance using standard classification metrics \- Provide an interactive user interface \- Demonstrate the application of machine learning to a Nigerian context \## ✨ Features \- Text-based language identification \- Supports English, Hausa, Igbo, and Yoruba \- TF-IDF character n-gram feature extraction \- Logistic Regression classification \- Accuracy, precision, recall, and F1-score evaluation \- Confusion matrix visualization \- Manual testing with unseen phrases \- Model confidence/probability estimates \- Interactive Gradio web interface \- Saved model and vectorizer for reuse \## 🧠 How It Works The application follows this pipeline: User enters text ↓ TF-IDF Character N-Gram Vectorization ↓ Logistic Regression Classifier ↓ Language Prediction ↓ Confidence / Probability Scores 1\. Text Input - The user provides a short phrase. Example: Bawo ni o se wa loni? 2\. Feature Extraction The …