Combines text, image, and metadata features using deep learning to capture cross-modal signals and context. This fusion improves the accuracy of fake news detection across diverse content types.
Uses NRC Emotion Lexicon and emoji semantics to detect public emotional response and novelty cues in replies.
Regions extracted from news titles are mapped to IDs, then encoded using neural networks to generate embeddings. These vectors capture geographic context for improved fake news detection.
Instantly displays a credibility label and sentiment score with color-coded feedback, helping users quickly assess the authenticity and emotional tone of news content.
The plugin integrates behavioural signals, text semantics, and image content to detect fake news in real-time while browsing social media posts.
Captures Reddit post IDs in real time as users scroll. Sends the data via an HTTP POST request to the FastAPI backend (/analyze endpoint).
Hosted on AWS EC2, the FastAPI backend receives requests and uses PRAW (Python Reddit API Wrapper) to fetch the full Reddit post content based on the post ID.
Fetched content is passed through multiple ML feature extractors such as text cleaning, emotion classification, and region detection, preparing the data for analysis.
The processed data is sent to a Hybrid-BERT trained model that returns a real/fake label and a confidence score. This result is returned as JSON and displayed directly inside the Reddit feed via the extension.
We use a Hybrid BERT model that combines deep language understanding with tabular features like sentiment, emotion, and engagement. This fusion improves fake news detection by capturing both context and behavior. Built with FastAPI and Reddit APIs, our system enables accurate, real-time analysis for social media platforms.
