基于多变量灰色预测模型的食品安全风险预警研究

The research about the constructions of multivariable grey forecasting model is proposed in this project. we apply the proposed model to solve the risk early warning of food safety. Research on multivariate prediction model, influence mechanism and rule number by parameter transformation technique for model and simulation prediction; genetic algorithm, particle swarm algorithm, wavelet analysis and optimization of multi variable grey prediction model based on the optimization of background value; considering the driving factors of the interaction, to construct interactive effects of multi variable grey forecast model and optimization; construction of modeling mechanism and dynamic time warping algorithm of multivariable non equidistance grey forecasting model; consider the delay phenomenon cause and effect between variables, the construction of driving multivariable grey prediction model and its optimization based on control factors. Finally, the prediction model proposed in this project is applied to the early warning research of food safety risk, which provides information and help for the relevant government departments.

本项目研究多变量灰色预测模型构建问题，并把提出的模型用于解决食品安全风险预警。研究多变量预测模型中，数乘变换技术对模型的参数和模拟预测值的影响机制与规律；基于遗传算法、粒子群算法、小波分析等优化多变量灰色预测模型背景值的优化问题；考虑驱动因素交互作用，构建具有交互效应的多变量灰色预测模型及其优化求解；采用动态时间规整算法实现多变量非等间距灰色预测模型的构建与建模机理研究；考虑变量之间因果作用的时滞现象，构建基于驱动因素控制项的多变量时滞灰色预测模型及其优化。最后，把本项目中提出的预测模型用于食品安全风险的预警研究，为政府相关部门决策提供信息与帮助。

本项目研究多变量灰色预测模型构建问题，并把提出的模型用于解决产业经济、能源经济、食品安全风险预警等问题解决中。 通过项目团队的研究，进行了数乘变换技术的研究、基于背景值优化的多变量灰色预测模型构建研究、考虑驱动变量控制项的多变量灰色预测模型构建研究、考虑时滞效应的多变量灰色预测模型研究等理论创新研究，并把提出的理论创新模型用于解决产业发展规划、园区发展规划、食品安全风险预警等相关问题中，取得了较好效果。多项研究结果为相关部门能源政策制定、食品安全风险预警等提供决策支持，促进了灰色预测理论创新与实践应用发展。