Fake News Detection using Machine Learning with TF-IDF (Word & Character n-grams) and SVM, Naive Bayes, and Logistic Regression classifiers. Includes preprocessing for multilingual (Tigrigna & English) text and model evaluation with cross-validation.
# Tigrigna Fake News Detection System
## Overview
The Tigrigna Fake News Detection System is a full-stack AI-powered web application developed to classify Tigrigna news content as either **REAL** or **FAKE** using Natural Language Processing (NLP) and Machine Learning techniques.
The project was developed as a final-year Computer Science and Engineering project with the goal of addressing misinformation challenges in low-resource languages such as Tigrigna.
The system integrates a React frontend, FastAPI backend, and machine learning models to provide real-time fake news prediction and analysis.
---
## Key Features
* Real-time fake news prediction
* Machine learning-based text classification
* Confidence score analysis
* Risk-level indication
* Prediction history management
* Responsive frontend interface
* REST API integration using FastAPI
* SQLite database integration
* Full-stack cloud deployment
=======
An AI-powered full-stack web application for detecting fake news in the
Tigrigna language using Natural Language Processing (NLP) and Machine Learning.
---
# π Project Overview
The **Tigrigna Fake News Detection System** is a full-stack intelligent web application developed to automatically classify Tigrigna news content as either:
- β
REAL
- β FAKE
using **Natural Language Processing (NLP)** and **Machine Learning** techniques.
The project was developed as a **final-year Computer Science and Engineering team project** to help combat misinformation in low-resource languages such as **Tigrigna**.
---
## Technologies Used
### Frontend
* React.js
* Vite
* Axios
* CSS
### Backend
* FastAPI
* Uvicorn
* SQLAlchemy
* Pydantic
* SQLite
### Machine Learning & NLP
* Scikit-learn
* Pandas
* NumPy
* Joblib
=======
- π° Real-time fake news prediction
- π€ AI-powered text classification
- π Confidence score analysis
- β οΈ Risk-level indication
- π Prediction history management
- π¨ Modern responsive React UI
- β‘ FastAPI backend API
- βοΈ Full-stack deplo β¦