Media
3 items
Description
Advanced fake review detection on Amazon using Aspect-Level Sentiment Analysis + Graph Neural Networks (97% accuracy) 🔍 Real-time fake review detection on Amazon with 97% accuracy. ALSAMIR (Aspect-Level Sentiment-Aware Multi-modal Inter-Review) is a powerful Chrome extension that automatically detects fake and misleading reviews on Amazon product pages. ⚡ HOW IT WORKS: - Automatically scans reviews when you open any Amazon product page - Extracts review text, ratings, helpful votes, and verified purchase status - Sends data to our machine learning model (XGBoost + all-MiniLM-L6-v2) - Analyzes 402 different features per review including text meaning, aspect sentiment, and behavioral patterns - Displays color-coded badges directly on each review 🎯 WHAT YOU SEE: - 🟢 Green badge with percentage = Genuine review - 🔴 Red badge with "FAKE X%" = Suspicious/fake review 📊 MODEL PERFORMANCE: - Accuracy: 97.07% - Precision: 94.01% - Recall: 91.35% - F1-Score: 92.66% 🔒 PRIVACY & SECURITY: - No personal data collected or stored - No cookies or tracking - Review texts are sent to our API only for detection and are not stored - HTTPS encryption for all communication 🛒 SUPPORTED PLATFORMS: - Amazon.in (India) - Amazon.com (USA) - More platforms coming soon (Flipkart, Meesho) 🆓 COMPLETELY FREE TO USE 👨🎓 Final Semester Project by Mahendra (MCA) 📚 Based on ALSAMIR paper (Knowledge-Based Systems, 2025) Install now and never be fooled by fake reviews again!