基于GA-XGBoost的宁波港物流需求预测  

Forecast of Ningbo Port Logistics Demand Based on GA-XGBoost

作　　者：李顺 李君[2] 吴鑫 梅碧舟 LI Shun;LI Jun;WU Xin;MEI Bi-zhou(Zhejiang Wanli University,Ningbo Zhejiang 315100;Zhejiang Yiduan Precision Machinery Co.,Ltd.,Xiangshan Zhejiang 315700)

机构地区：[1]浙江万里学院信息与智能工程学院,浙江宁波315100 [2]浙江万里学院,浙江宁波315100 [3]浙江易锻精密机械有限公司,浙江象山315700

出　　处：《浙江万里学院学报》2021年第2期71-77,共7页Journal of Zhejiang Wanli University

基　　金：宁波市科技厅惠民项目(2017C50028);国家级大学生创新创业训练计划项目(201910876035)。

摘　　要：通过灰色关联法,分析宁波港物流需求同腹地经济的相关度,建立宁波港物流需求预测指标集,提出一种基于遗传算法优化的极限梯度提升树模型。实验结果表明:这种港口物流需求预测模型取得的平均绝对误差和平均绝对百分比误差分别为21.62和1.05%,优于近年来的港口物流需求预测模型。Through grey correlation method,this paper analyzed the correlation between Ningbo port logistics demand and hinterland economy,established Ningbo port logistics demand forecasting index set,and put forward the eXtreme Gradient Boosting model based on genetic algorithm optimization.The results showed that the mean absolute error and the mean absolute percentage error obtained by this port logistics demand prediction model were 21.62 and 1.05%respectively,which were better than those of the port logistics demand prediction models in recent years.

关 键 词：港口物流 需求预测 极限梯度提升树 遗传算法 

分 类 号：F552.7[经济管理—产业经济]