IJRR

International Journal of Research and Review

| Home | Current Issue | Archive | Instructions to Authors | Journals |

Year: 2026 | Month: March | Volume: 13 | Issue: 3 | Pages: 327-333

DOI: https://doi.org/10.52403/ijrr.20260338

Multi-Modal System for Fake Review Detection on E-Commerce Platforms

Putti Venkata Siva Teja1, P. Purna Chandrika1, N. Lavanya Sai2, M. Sushma3, Md. Allabakshu4, T. Mahesh5

1,2,3,4,5Department of Information Technology,
Dhanekula Institute of Engineering & Technology, Vijayawada, Andhra Pradesh, India.

Corresponding Author: Putti Venkata Siva Teja

ABSTRACT

The expansion of e-commerce has made online reviews a primary factor in consumer purchasing decisions. However, the rise of fake reviews reduces trust in these platforms and creates confusion for buyers. This project develops a multi-modal detection system that analyzes both text and image-based reviews to identify fraudulent content. For textual data, we use Natural Language Processing (NLP) with the NLTK library and TF-IDF feature extraction to find patterns in deceptive writing. For images, Perceptual Hashing (pHash) is used to flag duplicate photos often used in fake reviews. Our system categorizes reviews using a Naive Bayes classifier and visualizes the results for specific products via bar graphs. Experimental results show high accuracy, helping consumers make more informed choices.

Keywords: Natural Language Processing, Machine Learning, Image Hashing, E-Commerce Systems, NL TK, TF-IDF.

[PDF Full Text]