基于SSA-XGBoost的综合型商业建筑停车需求预测研究  

Research on Parking Demand Forecast of Integrated Commercial Buildings Based on SSA-XGBoost

作　　者：李聪颖[1] 贠开拓 张浩星 张洪涛 袁锴璐 李坤[1] 吴佳西 LI Congying;YUN Kaituo;ZHANG Haoxing;ZHANG Hongtao;YUAN Kailu;LI Kun;WU Jiaxi(School of Civil Engineering,Xi’an University of Architecture and Technology,Xi’an 710055,China;New Energy Engineering Company,Guangdong Electric Power Design Institute,Guangzhou 510799,China)

机构地区：[1]西安建筑科技大学土木工程学院,西安710055 [2]中国能源建设集团广东省电力设计研究院,广州510799

出　　处：《武汉理工大学学报(交通科学与工程版)》2025年第1期15-20,27,共7页Journal of Wuhan University of Technology(Transportation Science & Engineering)

基　　金：陕西省自然科学基础研究计划项目(2020JM-478)。

摘　　要：文中基于综合型商业建筑停车需求与机动车吸引量的关系,构建综合型商业建筑停车需求影响因素体系;运用麻雀搜索算法优化极限梯度提升树的超参数,建立综合型商业建筑停车需求预测组合模型;以西安市58个综合型商业建筑的停车需求预测为例,对比SSA-XGBoost模型与支持向量回归模型、XGBoost模型、lasso回归模型的预测结果.结果表明:SSA-XGBoost模型的R2值为0.963、平均绝对误差为75.584、均方根误差为85.749,相较于其他几种预测模型有更高的R2值和更小的预测误差.Based on the relationship between parking demand of comprehensive commercial buildings and motor vehicle attraction,the influencing factors system of parking demand of comprehensive commercial buildings is constructed.The sparrow search algorithm is used to optimize the superparameter of the limit gradient lifting tree,and a comprehensive parking demand forecasting combination model of commercial buildings is established.Taking the parking demand prediction of 58 comprehensive commercial buildings in Xi’an as an example,the prediction results of SSA-XGBoost model,support vector regression model,XGBoost model and lasso regression model are compared.The results show that the R 2 value of SSA-XGBoost model is 0.963,the average absolute error is 75.584,and the root mean square error is 85.749.Compared with other prediction models,SSA-XG Boost model has higher R 2 value and smaller prediction error.

关 键 词：停车需求预测 综合型商业 XGBoost 麻雀搜索算法 组合模型 

分 类 号：U491.7[交通运输工程—交通运输规划与管理]