• Ajitabh Soni
• Nirbhay Singh

Online shopping platforms increasingly rely on user-generated reviews to guide consumer decisions. However, the growing prevalence of fake or manipulated reviews threatens trust, misguides customers, and damages seller reputations. While existing detection methods have advanced from classical machine learning (ML) to deep learning (DL) and hybrid approaches, most remain black-box systems with limited real-world deployment. This paper proposes a real-time explainable fake review detection system that integrates multi-model ML and NLP techniques with explainability modules such as LIME. Unlike prior systems, our solution provides transparent justifications for classifications, helping users understand why a review is deemed genuine or fake. The system supports dual input—product URLs or direct review text—via an Oxylabs-powered scraper and is deployed as a full-stack application using Flask (backend) and ReactJS (frontend) and MongoDB (database). Experimental analysis with multi-domain datasets demonstrates that our hybrid model outperforms baseline classifiers in both accuracy and interpretability. The research contributes towards building trustworthy, real-time, and user-friendly fake review detection systems, addressing critical gaps in e-commerce fraud prevention.

Fake reviews, E-commerce, NLP, Machine learning, Explainable AI, Deep learning, Database, Trust, Transparency.

"REAL TIME EXPLAINAINABLE FAKE REVIEW DETECTION IN E-COMMERCE USING MACHINE LEARNING AND NLP ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 9, page no.e439-e442, September-2025, Available :http://www.jetir.org/papers/JETIR2509449.pdf

"REAL TIME EXPLAINAINABLE FAKE REVIEW DETECTION IN E-COMMERCE USING MACHINE LEARNING AND NLP ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 9, page no. ppe439-e442, September-2025, Available at : http://www.jetir.org/papers/JETIR2509449.pdf