BP神经网络在水产品安全风险预警中的应用

Application of BP Neural Network to Safety Risk Forewarning

近年来,消费者对水产品的安全和品质提出了更高的要求,而水产品质量安全问题可能发生在其供应链的各个环节,水产品质量一旦出了问题,既危害消费者的健康,又给企业带来经济损失。针对上述问题,通过分析水产品供应链中安全隐患的因素,整合供应链上的追溯信息和监测信息,建立水产品预警指标体系。采用BP神经网络搭建安全预警模型,并对模型进行训练和预测,输入、分析预警和水产品添加剂指标数据,确定预警警度,最终确定供应链安全风险等级。预警结果表明,该方法较粗糙集理论、时序分析、模糊综合评价、回归分析等方法,其解决实际问题中预测误差小,可有效提高水产品供应链风险预警的准确性。

Recent years, consumers have more demand for the safety and quality of aquatic products. The quality and safety problems will occur everywhere in the supply chain. Once there is a quality and safety problem, not only the health of consumers will he endan- gered but also the company will have financial losses. To solve these problems, we analyze the factors which affect the safety and quali- ty in the aquatic product chain, and integrate tracking information and monitor information in the supply chain, and then establish the forewarning index system of aquatic products. We build the forewarning model by using the BP neural network, and train it using a great of data. The warning degree depends on the analysis of the input data and aquatic products additive index data, and confirm the degree of risk. The results show that this method has fewer error than rough set theory, time - series, fuzzy comprehensive evaluation, regression analysis and other methods when solving practical problems, and can effectively improve the accuracy of the risk forewarning for supply chain.