首页|基于Adaboost回归算法的安徽省物流需求短期预测研究

基于Adaboost回归算法的安徽省物流需求短期预测研究

扫码查看
[目的]物流需求预测有助于调整物流资源的分配,对促进物流业发展具有重要意义.[方法]选取安徽省1995-2022年与物流需求相关的指标数据为原始样本数据,用货运量来表征物流需求.通过XGBoost特征选择算法筛选出6个用于预测的指标.在此基础上,使用3种方法分别构建模型,并对这些模型进行对比分析.最终,选择精度最高的Adaboost回归算法来预测安徽省短期物流需求.[结果]2023-2026 年,安徽省的物流需求预测值分别为 402 942.428 万t、369 877.222万t、380 884.375万t、382 319.5万t.[结论]未来四年,安徽省物流的货运量呈不稳定发展态势.根据安徽省的区位优势及疫情的全面开放,安徽省物流业表现出较大的发展潜力.
Research on Short-term Forecasting of Logistics Demand in Anhui Province Based on Adaboost Regression Algorithm
[Purposes]Logistics demand forecasting is helpful to adjust the distribution of logistics re-sources and is of great significance to promote the development of logistics industry.[Methods]The index data related to logistics demand in Anhui Province from 1995 to 2022 were selected as the original sample data,and the freight volume was used to characterize the logistics demand.Six indicators for prediction were selected by XGBoost feature selection algorithm.On this basis,three methods are used to construct models respectively,and these models are compared and analyzed.Finally,the Adaboost regression algorithm with the highest accuracy is selected to predict the short-term logistics demand of Anhui Province.[Findings]From 2023 to 2026,the predicted values of logistics demand in Anhui Province were 4029.42428 million tons,3698.77222 million tons,3808.84375 million tons and 3823.195 million tons,respectively.[Conclusions]In the next four years,the freight volume of logistics in Anhui Province will show a fluctuating.Based on the geographical advantages of Anhui Province and the comprehensive loosening restrictions of the epidemic,the logistics industry in Anhui Province has shown a lot of devel-opment potential.

Adaboostfeature selectionlogistics demand forecastAnhui province

荀守奎、葛成丽

展开 >

安徽理工大学经济与管理学院,安徽 淮南 232001

安徽理工大学金融科技研究所,安徽淮南 232001

安徽理工大学城镇化与产业发展研究所,安徽 淮南 232001

Adaboost 特征选择 物流需求预测 安徽省

安徽省教育厅人文社科重点项目安徽省教育厅2021年度质量工程煤炭行业教育研究课题2022年度安徽省级新时代育人质量工程(研究生教育)项目

SK2019A01002021sx0332021MXJG0512022szsfkc075

2024

河南科技
河南省科学技术信息研究院

河南科技

影响因子:0.615
ISSN:1003-5168
年,卷(期):2024.51(2)
  • 15