Enterprise Supply Chain Risk Assessment Based on the Support Vector Machine Algorithm and Fuzzy Model

Abstract

In systems with uncertain information, ambiguity must be taken into account. In this paper, fuzzy set theory concepts are incorporated into support vector machines (SVM). This ensemble preserves the benefits of SVM regression models and fuzzy regression models, where SVM learning theory describes the properties of learning machines that enable them to generalize well, and fuzzy set theory offers an efficient method for capturing the approximate, imprecise properties of the real world. In accordance with the phase space reconstruction theory of dynamical systems, a fuzzy model for enterprise supply chain risk assessment is proposed using the robust nonlinear mapping capability of support vector machines and the characteristic of fuzzy logic that makes it easy to combine prior system knowledge into fuzzy rules. The results demonstrate that the prediction model can not only automatically acquire knowledge from learning data to generate fuzzy rules but can also extract support vectors that can represent the inherent laws of enterprise supply chain risks, drastically reduce the number of support vectors, and accurately predict. Future risk can be accurately predicted. This conclusion suggests that the support vector machine based on a fuzzy model is an effective method for analyzing enterprise supply chain risk.