DominionAkinrotimi/Yoruba-Offensive-Language-Detection-Model

Detection of Offensive Language in Yoruba Text using ML (TF-IDF, Logistic Regression) # Offensive Yoruba Language Detection App This repository contains a machine learning-based application for detecting offensive language in yoruba text. The app is built using either **Streamlit** or **Flask** and leverages a **Logistic Regression** model trained on a dataset of tweets. The model uses **TF-IDF Vectorization** for text preprocessing and classification. --- ## Table of Contents 1. Project Overview 2. Features 3. Installation 4. Usage 5. Model Development 6. File Structure 7. Contributing 8. License --- ## Project Overview The goal of this project is to detect if a lanuange in yoruba is based off of text inputs. The app takes a sentence as input and predicts whether it contains offensive language, hate speech, or is normal. The model is trained on a dataset of tweets and uses **TF-IDF Vectorization** for feature extraction and **Logistic Regression** for classification. --- ## Features - **Text Input**: Users can input a sentence to check for offensive language. - **Real-Time Prediction**: The app provides instant predictions using a pre-trained machine learning model. - **Clean and Preprocess Text**: The app cleans and preprocesses the input text (e.g., removes emojis, URLs, and special characters) before making predictions. - **Streamlit and Flask Support**: The app can be deployed using either Streamlit or Flask. --- ## Installation ### Prerequisites - Python 3.7 or higher - pip (Python package manager) ### Steps 1. **Clone the Repository**: ```bash git clone github.com cd Yoruba-Offensive-Language-Detection-Model ``` 2. **Create a Virtual Environment** (Optional but Recommended): ```bash python -m venv venv source venv/bin/activate # On Windows: venv\Scripts\activate ``` 3. **Install Dependencies**: ```bash pip install -r requirements.txt ``` 4. **Download NLTK Data**: The app uses NLTK for text preprocessing. Download the required NLTK data by running: ```b …