Year: 2007
Type: article
Abstract: People often enmesh the Internet auction frauds which damage the benefits of Internet market and threaten transactions security. This research applies social network analysis and data mining to extract characteristic features from two random collected transaction datasets of Yahoo auction site. One dataset is used to construct prediction model and another is treated as validation. The average accuracy ratio of proposed model is at least 90%. The findings are: (1) the abnormal accounts involve circular transaction; (2) fraud accounts can accumulate higher positive reputations in very short time from its circular transaction and rarely play key nodes in transaction network.
Authors: Yungchang Ku, Chaochang Chiu
Language: English
Topic: Spam and Phishing Detection
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W3012473159
