基于XGBoost国省干线沥青路面使用性能预测研究  

Research on Service Performance Prediction of Asphalt Pavement Based on XGboost National and Provincial Trunk Line

作　　者：侯美晴 粟海涛 杨洋[1] 亓祥宇 刘名扬 HOU Meiqing;SU Haitao;YANG Yang;QI Xiangyu;LIU Mingyang(Key Laboratory of Disaster Prevention and Mitigation,School of Civil Engineering,Kunming University of Science and Technology,Kunming 650000,China;Yunnan Provincial Institute of Highway Science and Technology,Kunming 650000,China)

机构地区：[1]昆明理工大学建筑工程学院防灾减灾重点实验室,昆明650000 [2]云南省公路科学技术研究院,昆明650000

出　　处：《武汉理工大学学报（交通科学与工程版）》2024年第3期574-578,583,共6页Journal of Wuhan University of Technology（Transportation Science & Engineering）

基　　金：国家自然科学基金项目(11962009);云南省交通运输厅科技创新及示范项目(云交科教便[2021]86号-(三))。

摘　　要：文中提出了一种基于XGBoost(extreme gradient boost)算法的国省干线沥青路面损坏状况指数(pavement condition index,PCI)预测模型,并与传统回归模型预测结果进行对比.结果表明:通过对比分析不同模型的3个定量评价指标(均方根误差、平均绝对误差和拟合系数)以及最终预测结果的折线图,证明采用考虑多因素条件的XGBoost算法PCI预测模型的预测精度要优于传统回归预测模型,验证了该模型的有效性和优越性.A prediction model of pavement condition index(PCI)of national and provincial trunk lines based on XGBoost(extreme gradient boost)algorithm was proposed.Compared with the traditional regression model,the results show that by comparing and analyzing three quantitative evaluation indexes(root mean square error,average absolute error and fitting coefficient)of different models and the line chart of the final prediction results,it is proved that the prediction accuracy of PCI prediction model with XGBoost algorithm considering multi-factor conditions is better than that of traditional regression prediction model,and the effectiveness and superiority of this model are verified.

关 键 词：道路工程 沥青路面 路面性能预测 路面状况指数(PCI) 机器学习XGBoost 

分 类 号：U418.6[交通运输工程—道路与铁道工程]