FakeThreads Investigating Fake News Dissemination Patterns in Threads

The rapid proliferation of misinformation on social media platforms poses significant challenges to information integrity and public discourse. This project focuses on developing a binary classification system to detect fake threads (conversations containing misleading or false information) on Threads. We employed traditional machine learning models, including Random Forest and LightGBM, with TF-IDF vectorization for feature extraction, alongside advanced transformer-based approaches using DistilBERT and GPT-2 for contextual embeddings. Each model was rigorously evaluated using standard metrics such as accuracy, precision, recall, and F1-score to compare performance. To enhance user accessibility, a local web interface was built, featuring user registration and login functionality. Authenticated users can input thread text, receiving real-time predictions of fake or genuine classification, accompanied by LIME explanations highlighting influential words contributing to the decision. This interpretable system aids in understanding fake news dissemination patterns and promotes transparency in detection efforts.

Keywords

Fake news detection, Threads platform, Misinformation, Binary classification, Random Forest, LightGBM, TF-IDF, DistilBERT, GPT-2, LIME explainability, Web interface, Machine learning, Social media analysis