Fake News Detection using LSTM

This repository contains the code and report for a fake news detection project using Long Short-Term Memory (LSTM) neural networks. The model was implemented in Python using TensorFlow and Keras and was designed to classify news articles as either real or fake.

Introduction

This project was made to address the current problem of misinformation by training a machine learning model that is capable of indentifying fake news from news articles. Results

Results

The model achieved:

Accuracy: ~93%

Precision: 92%

Recall: 93%

F1-Score: 0.92 (real news) and 0.93 (fake news)

Graphs of accuracy and loss across epochs showed consistent learning with minor fluctuations.

References

The data was taken from kaggle: Fake News Detection Datasets