Fraudulent E-Commerce Websites Detection Through Machine Learning
-
1
Universidad de León
info
- 2 INCIBE (Spanish National Cybersecurity Institute, León)
- Hugo Sanjurjo González (coord.)
- Iker Pastor López (coord.)
- Pablo García Bringas (coord.)
- Héctor Quintián (coord.)
- Emilio Corchado (coord.)
Éditorial: Springer International Publishing AG
ISBN: 978-3-030-86271-8, 978-3-030-86270-1
Année de publication: 2021
Pages: 267-279
Congreso: Hybrid Artificial Intelligent Systems (HAIS) (16. 2021. Bilbao)
Type: Communication dans un congrès
Résumé
With the emergence of e-commerce, many users are exposed to fraudulent websites, where attackers sell counterfeit products or goods that never arrive. These websites take money from users, but also they can stole their identity or credit card information. Current applications for user protection are based on blacklists and rules that turn out into a high false-positive rate and need a continuously updating. In this work, we built and make publicly available a suspicious of being fraudulent website dataset based on distinctive features, including seven novel features, to identify these domains based on recently published approaches and current web page properties. Our model obtained up to 75% F1- Score using Random Forest algorithm and 11 hand-crafted features, on a 282 samples dataset.