Loading article...
Loading article...
Stream/Area:
-
Subject:
-
Paper ID:
IJSREM66638
Volume:
10 |
8 |
Aug 2026
Pages:
1-5
Publication Frequency:
Monthly (12 issues per year)
License:International Journal of Scientific Research in Engineering and Management follows the Creative Commons Attribution 4.0 (CC BY 4.0) License, allowing free use and distribution with proper attribution.
Artificial Intelligence and Data Science
Avinashilingam Institute for Home Science and Higher Education for Women
Artificial Intelligence and Data Science
Avinashilingam Institute for Home Science and Higher Education for Women
Artificial Intelligence and Data Science
Avinashilingam Institute for Home Science and Higher Education for Women
Artificial Intelligence and Data Science
Avinashilingam Institute for Home Science and Higher Education for Women
Artificial Intelligence and Data Science
Avinashilingam Institute for Home Science and Higher Education for Women
E-commerce platforms increasingly rely on discount-based promotions to influence consumer purchasing decisions, particularly during major sale events. This reliance has given rise to a widespread and largely undetected form of pricing fraud, in which sellers artificially inflate the Maximum Retail Price (MRP) prior to a sale and subsequently apply a discount, creating the illusion of substantial savings that do not reflect any genuine price reduction. Existing price-tracking tools such as Keepa and CamelCamelCamel provide descriptive historical price data but lack the diagnostic capability to determine whether a discount is genuine. This paper presents FraudLens, a hybrid, data-driven fraud detection framework that combines an unsupervised anomaly detection layer (Isolation Forest) with supervised classification models (Random Forest and XGBoost) to identify manipulated discounts across Amazon and Flipkart product listings. The system incorporates domain-driven feature engineering — including MRP inflation ratio, discount permanence score, pre-sale price spike detection, and seller trust score — together with SHAP-based explainability to translate model predictions into human-readable justifications. The proposed system was trained and validated on a combined dataset of 73,114 records spanning 3,000 unique products, and the XGBoost classifier achieved a weighted F1-score of 96.9%, correctly classifying products as Genuine, Suspicious, or Fake. An interactive Streamlit dashboard was also developed to provide real-time, per-product fraud verdicts supported by visual evidence, enabling consumers to make more informed purchasing decisions.
KEYWORDS
This project has designed, implemented, and validated FraudLens, a hybrid fraud detection framework that combines rule-based logic, unsupervised anomaly detection, and supervised machine learning classification to identify fake discounts in e-commerce. By anchoring every discount claim against a 90-day historical price baseline and layering Isolation Forest anomaly scores into a Random Forest and XGBoost classification pipeline, the system achieves high detection accuracy (weighted F1-score of 96.9%) while remaining interpretable through SHAP-based explainability. The accompanying interactive dashboard demonstrates that fraud detection results can be delivered to consumers in a transparent, actionable, and easy-to-understand format, moving beyond descriptive price trackers toward a genuine diagnostic tool for pricing integrity.
[1] H. L. S. Gino, J. R. Ponciano, C. D. G. Linhares, and A. J. M. Traina, “ECommVis: Supporting E-Commerce Marketplace Advertising Outcomes Through a Visual Analytics System,” IEEE Open Journal of the Computer Society, vol. 7, pp. 455–467, 2026. [2] Y. Yada, J. Feng, T. Matsumoto, N. Fukushima, F. Kido, and H. Yamana, “Dark Patterns in E-Commerce: A Dataset and Its Baseline Evaluations,” in Proc. 2022 IEEE International Conference on Big Data (IEEE BigData), Osaka, Japan, 2022, pp. 3015–3022, doi: 10.1109/BigData55660.2022.10020800. [3] J. Ramakrishnan, E. Shaabani, C. Li, and M. A. Sustik, “Anomaly Detection for an E-Commerce Pricing System,” arXiv preprint arXiv:1902.09566, 2019. [4] A. Mathur, G. Acar, M. J. Friedman, E. Lucherini, J. Mayer, M. Chetty, and A. Narayanan, “Dark Patterns at Scale: Findings from a Crawl of 11K Shopping Websites,” Proc. ACM Hum.-Comput. Interact., vol. 3, no. CSCW, Article 81, pp. 1–32, Nov. 2019, doi: 10.1145/3359183. [5] Y. Xu, M. Li, L. Cui, S. Huang, F. Wei, and M. Zhou, “LayoutLM: Pre-training of Text and Layout for Document Image Understanding,” in Proc. 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (KDD ’20), 2020, pp. 1192–1200, doi: 10.1145/3394486.3403172.
Publication Date
8/10/2026
Peer Review Type
Double-Blind Peer Review
Review Rounds: 2
Documents
Article Views
0
Total views since publication
PDF Downloads
2
Total downloads since publication
Geographic Reach
5
Digital Preservation: This article is permanently archived on Zenodo and complies with all data preservation standards across major indexing platforms.
Plagiarism Screening
This manuscript has been checked for plagiarism using TurnitinPlagiarismChecker software.
Conflict of Interest
The authors declare no conflict of interest related to this research work.
Funding Information
No funding information available for this research.
Related research in the same subject area
Loading similar articles...