E-Commerce Fraud Detection Model by Computer Artificial Intelligence Data Mining.

Area of Science:

• Computer Science
• Data Science
• Financial Technology

Background:

• E-commerce enterprises face significant financial risks due to fraudulent activities.
• Existing methods for fraud detection have limitations in accuracy and data integration.
• Big Data Mining (BDM) offers potential for analyzing large datasets to identify risks.

Purpose of the Study:

• To identify e-commerce fraud and mitigate financial risks using advanced technologies.
• To develop an effective e-commerce Fraud Detection Model (FDM) leveraging Information Fusion Technology (IFT).
• To compare the performance of the proposed FDM against established models like Support Vector Machine (SVM) and Logistic Regression Model (LRM).

Main Methods:

• Utilized Big Data Mining (BDM) for risk analysis.
• Developed an IFT-based FDM integrating Computer Technology (CT), Artificial Intelligence (AI), and Data Mining (DM).
• Compared the proposed FDM with SVM and LRM for e-commerce fraud risk assessment.

Main Results:

• The IFT-based FDM demonstrated significantly higher fraud identification accuracy for Business-to-Business (B2B) e-commerce enterprises compared to SVM and LRM.
• The proposed FDM effectively fuses diverse information sources for comprehensive risk analysis.
• The model accurately analyzes enterprise financial and credit status to predict fraudulent behavior probability.

Conclusions:

• The IFT-based FDM is superior to SVM and LRM in processing and calculating financial risk data from multiple sources.
• The developed FDM provides a robust method for B2B e-commerce fraud identification.
• This research offers technical support for the healthy development of e-commerce by mitigating fraud risks.