基于Holt-Winters和PSO-SVR的快递业务量最优加权组合预测  

Optimal Weighted Combination Forecast of Express Delivery Volume Based on Holt-Winters and PSO-SVR

作　　者：周晓晔[1] 成佳慧 戴思聪 ZHOU Xiao-ye;CHENG Jia-hui;DAI Si-cong(School of Management,Shenyang University of Technology,Shenyang 110870,China)

机构地区：[1]沈阳工业大学管理学院,辽宁沈阳110870

出　　处：《数学的实践与认识》2023年第7期51-60,共10页Mathematics in Practice and Theory

基　　金：辽宁省社会科学规划基金重大委托项目(L22ZD010)。

摘　　要：实现快递业务量的精准预测意义重大,为了进一步提高快递业务量的预测精度,提出一种基于霍尔特-温特斯(Holt-Winters)和粒子群优化支持向量回归(PSO-SVR)的最优加权组合预测模型.首先利用最优加权组合预测方法将构建的Holt-Winters模型和PSO-SVR模型进行组合,然后选取我国2017-2021年快递业务量季度数据验证模型的有效性,将组合模型与GM(1,1)等经典模型的预测结果相比较,最后对我国2022年第1季度至2023年第2季度的快递业务量进行预测.结果表明,组合模型得出的预测值更准确,预测精度优于对比模型,具有一定的理论和实用价值.It is of great significance to realize the accurate prediction of express business volume.In order to further improve the prediction accuracy of express business volume,an optimal weighted combination prediction model based on Holt-Winters and particle swarm optimization support vector regression(PSO-SVR)is proposed.Firstly,the Holt-Winters model and PSO-SVR model are combined by using the optimal weighted combination forecasting method.Then the quarterly data of China's express business volume from 2017-2021 are selected to verify the effectiveness of the model,and the combined model is compared with the prediction results of classical models such as GM(1,1).Finally the prediction of express business volume from the 1st quarter of 2022 to the 2nd quarter of 2023 in China is made.The results show that the combined model is more accurate and the prediction accuracy is better than the contrast models,which has a certain theoretical and practical value.

关 键 词：Holt-Winters模型 PSO-SVR模型 最优加权组合预测 快递业务量预测 

分 类 号：F259.2[经济管理—国民经济] F224