摘要
为解决边境封控油料保障需求预测问题,针对油料保障需求不仅呈现线性变化,且蕴含非线性变化规律,传统的单一模型很难同时对线性和非线性规律加以预测的特点,提出一种基于灰色支持向量机(GM-SVM)的边境封控油料保障需求组合预测模型。首先,运用灰色模型对油料保障需求进行预测,挖掘其线性变化规律;然后,采用支持向量机进行预测,描述油料保障需求的非线性变化规律;最后,将两种预测结果进行加权平均,作为边境封控油料保障需求预测的最终结果。预测结果表明,GM-SVM组合预测模型预测精度较高,较好地克服了单一预测模型的缺陷。
In order to solve the problem of demand forecast for POL support in border blockage and control,considering that POL support demand has the characteristic of linear variation and nonlinear variation and traditional single model couldn＇t forecast linear and nonlinear variation simultaneously,the paper puts forward a demand forecast model for POL support in border blockage and control based on GM-SVM. Firstly,it forecasts the POL support demand with grey model and excavates its linear variation rule. Then,it forecasts POL support demand with SVM and describes its nonlinear variation rule. Finally,it takes a weighted average on the two forecast results and obtains the final demand forecast result for POL support in border blockage and control. The result shows that the forecast model based on GM-SVM has high forecast accuracy and can overcome shortcoming of single forecast model.
出处
《军事交通学院学报》
2016年第3期90-94,共5页
Journal of Military Transportation University
关键词
油料保障
需求预测
灰色模型
支持向量机
POL support
demand forecast
grey model（GM）
support vector machine（SVM）
