E-Commerce Credit Risk Assessment Based on Fuzzy Neural Network.

Area of Science:

• Network science and artificial intelligence
• Financial risk management
• Supply chain finance

Background:

• Traditional network reconstruction methods often lack dynamic adaptability and efficient parameter optimization.
• Accurate credit risk assessment is crucial for the stability and growth of online supply chain finance.

Purpose of the Study:

• To propose a cooperative strategy-based self-organization mechanism for network reconstruction.
• To develop an efficient parameter optimization strategy for SC-IR2FNN (Self-adaptive Credit Information Risk Fuzzy Neural Network).
• To identify and analyze key factors influencing credit risk in online supply chain finance.

Main Methods:

• A comprehensive evaluation algorithm and structure adjustment mechanism for self-organization.
• A multilayer optimization engine for joint optimization of nonlinear and linear parameters in SC-IR2FNN.
• Principal component factor analysis and logistic regression to identify significant credit risk factors.
• An improved credit scoring system for online supply chain finance risk assessment.

Main Results:

• The proposed self-organization mechanism enables simultaneous network reconstruction and parameter optimization without pre-set thresholds.
• The multilayer optimization engine reduces computational complexity and improves the learning rate of SC-IR2FNN.
• Profitability, solvency, core enterprise profitability, operational guarantee, growth ability, supply chain online degree, financing enterprise quality, and cooperation factors are identified as key indicators for credit risk.
• Logistic model confirms financing company profitability, debt repayment, and core company profitability as significant impact factors.
• The improved credit scoring system effectively evaluates credit risk and addresses speculative and circular credit fraud.

Conclusions:

• The cooperative strategy-based self-organization mechanism offers an effective approach for adaptive network reconstruction.
• The developed parameter optimization strategy enhances the performance and efficiency of SC-IR2FNN.
• The study provides a robust framework for credit risk assessment in online supply chain finance, contributing to more secure financial ecosystems.