考虑整体和极端偏差的体育用品价格组合预测方法  

Sports Goods Price Combination Forecasting Method Considering the Whole And Extreme Deviation

作　　者：王泽 刘世超 陈曦[1] WANG Ze;LIU Shi-chao;CHEN Xi(College of Physical Education,Taiyuan University of Technology,Jinzhong 030600,China;Zhengzhou Food Engineering Vocational College,Zhengzhou 450000,China)

机构地区：[1]太原理工大学体育学院,山西晋中030600 [2]郑州食品工程职业学院,河南郑州450000

出　　处：《数学的实践与认识》2025年第3期25-38,共14页Mathematics in Practice and Theory

基　　金：山西省哲学社会科学规划课题(2023YY054)。

摘　　要：为解决传统预测模型存在模型单一、考虑片面等缺陷,提出一种考虑整体和极端偏差的组合预测方法.首先定义了描述数据尾部极端偏差的衡量指标:在险预测PaR和模型预期损失MES;然后构造考虑模型整体偏差与极端偏差的优化函数,并选择改进烟花算法(FWA)进行求解;最后通过在多解释变量数据的体育娱乐用品类城市商品零售价格指数预测和单条时序数据的体育娱乐用品类农村商品零售价格指数预测问题上进行验证.实验表明:本文所提的组合模型能够充分挖掘数据信息,解决数据过拟合问题,较数学算子法、权重分配法和信息法等基准预测模型性能更优.In order to solve the defects of traditional forecasting model,such as single model and one-sided consideration,a combined forecasting method considering the whole and extreme deviation is proposed.Firstly,a measure describing the extreme deviation in the data tail is defined:Prediction at Risk(PaR)and Model Expected Shortfall(MES).Then,the optimization function considering the overall deviation and extreme deviation of the model is constructed,and the improved fireworks algorithm(FWA)is selected to solve it.Finally,the paper verifies the prediction of urban retail price index of sports and entertainment goods with multi-explanatory variable data and the prediction of rural retail price index of sports and entertainment goods with single time series data.The experimental results show that the combined model proposed in this paper can fully mine the data information and solve the problem of data overfitting,and has better performance than the mathematical operator method,weight allocation method and information entropy method.

关 键 词：体育用品 组合预测 极端偏差 优化 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] F224[自动化与计算机技术—计算机科学与技术]