Automated Fake News Detection using Machine Learning Algorithms

This research project presents the development and comparative analysis of automated fake news detection systems using both traditional machine learning algorithms and advanced deep learning techniques. With the exponential growth of digital media, the spread of fake news has become a critical global issue, impacting public perception, political stability, and societal trust.

The study leverages a publicly available fake news dataset sourced from Kaggle, comprising both real and fake news articles. Rigorous data preprocessing techniques were applied, including text cleaning, tokenization, and feature extraction, to prepare the data for analysis.

The project implements and evaluates the performance of the following models:

Logistic Regression

Naive Bayes

Random Forest

Long Short-Term Memory (LSTM) neural network

Key performance metrics such as accuracy, precision, recall, F1-score, and ROC-AUC score were used to assess the models. Visualizations, including confusion matrices and ROC curves, provide a comprehensive understanding of each model's effectiveness.

Findings from this research indicate that while traditional machine learning models perform reasonably well, the LSTM model outperforms them by effectively capturing complex patterns and contextual dependencies in textual data. The project also includes WordCloud visualizations and performance comparison charts to aid interpretability.

All source code, dataset, visual assets, and the final research paper are included in this repository to promote transparency, reproducibility, and further research. This work aims to assist researchers, fact-checkers, and policy-makers in developing scalable solutions for combating misinformation in the digital age.

Keywords: Fake News, Machine Learning, Deep Learning, LSTM, Natural Language Processing, Text Classification, Data Science, Misinformation Detection, Research Paper, Open Science