摘要
在不确定环境下,区间数是复杂系统一种信息表达形式.论文在单项区间预测的基础上,提出单项预测方法与实际值序列之间的灰色关联度作为精度的度量准则,构建区间型组合预测模型.考虑模型求解的复杂性,从对策论的角度出发,把组合预测方法视为一个合作对策,单项预测方法视为合作对策的局中人,灰色关联度作为度量合作对策的收益函数,按照合作对策中Shapley值法在各单项预测模型中进行分配,从而给出组合预测权系数确定的新思路.为了验证模型的有效性,利用西德克萨斯中质原油现货价格,针对论文提出的组合模型进行比较分析,计算结果验证了论文提出的模型的可行性与有效性.
In uncertain environment,interval number was a form of information expression in complex system.Based on the single interval prediction,the grey correlation degree between the single interval prediction method and the actual value sequence was proposed as the measurement criterion of accuracy,and the interval combination prediction model was constructed.Considering the complexity of solving the model,from the perspective of game theory,the combination prediction method was regarded as a cooperative game,the single prediction method was regarded as a player of the cooperative game,the grey correlation degree was acted as the benefit function to measure the cooperative game,and the distribution was carried out in each single prediction model according to the Shapley value method in the cooperative game.A new way to determine the combination prediction weight coefficient was given.In order to show the validity of the model,the spot price of West Texas Intermediate crude oil was used to compare and analyze the combined model proposed in the paper.The calculation results indicated the feasibility and effectiveness of the model.
作者
陈勤勤
陈华友
韩冰
CHEN Qinqin;CHEN Huayou;HAN Bing(School of Big Data and Statistics,Anhui University,Hefei 230601,China;School of Mathematical Sciences,Anhui University,Hefei 230601,China)
出处
《安徽大学学报（自然科学版）》
CAS
北大核心
2023年第4期16-24,共9页
Journal of Anhui University(Natural Science Edition)
基金
国家自然科学基金资助项目(71871001)。
