E-Commerce across Boarder Logistics Risk Evaluation Model Based on Improved Neural Network

Abstract

BP neural network is a typical algorithm in artificial intelligence network. It has strong nonlinear mapping ability and is the most prominent part to solve some nonlinear problems. In the traditional BP algorithm, the coincidence initialization of weights and thresholds is random, which reduces the efficiency of the algorithm on the one hand and affects the accuracy of the algorithm results on the other hand. In order to solve these problems, this paper studies an e-commerce cross-border logistics risk assessment model based on improved neural network. This model can help merchants engaged in cross-border e-commerce to select appropriate third-party settlement platforms, so as to reduce the cost of merchants in the process of capital settlement. The key information in BP neural network algorithm is stored in weights and thresholds, which is enough to prove the importance of weights and thresholds for the effective operation of the whole network. The e-commerce cross-border logistics risk assessment model based on improved neural network aims to solve the problem of low level of risk assessment and the bottleneck of logistics risk assessment. The improved e-commerce cross-border logistics risk assessment model based on neural network can be used for risk rating before business development, so as to adopt different risk management methods for different risk levels.