期刊文献+

基于XGBoost的高速公路工程概算预测方法研究 被引量:5

Study on Prediction Method of Expressway Project Estimate Based on XGBoost
原文传递
导出
摘要 我国长期以来,高速公路工程概算的编制是基于定额体系进行编制,而定额体系的确定具有滞后性、偏差性的缺点,导致所编制的高速公路工程概算存在偏差,而越来越大的投资规模和承包模式,对于投资资金的管控也提出了更高的要求。为解决该问题,经过文献分析与方法研究,基于XGBoost算法模型与造价指标体系对高速公路工程概算进行了预测模型的创建,分析了XGBoost算法模型与指标体系的优势。以大量筛选指标和项目概算信息为样本库,使用网格搜索和交叉验证的方式优化了XGBoost算法模型,调整了模型相关的参数组合,优化了XGBoost算法模型的参数设置。以RMSE,MAE,R2这3个常用预测精度评价指标组合评价作为评价体系,对模型的预测从精准性、稳定性和偏差性上进行了评价。最终提出了造价预测的新方案,并对所创建的XGBoost算法预测模型进行了抽样验证,展现该算法模型在高速公路工程概算预测过程中的预测能力,对推广造价指标体系、提高高速公路工程造价预测的精度具有一定的促进作用。同时也对造价指标的影响因素进行了详细分析,以特征重要性为指标进行分析与探究,探究对于我国高速公路造价指标的重要影响因素,探究影响因素与造价指标之间的潜在关系,为投资方和设计单位提供概算编制的新的辅助参考指标和研究方向。 For a long time in China,the compilation of expressway project estimates is based on a quota system,and the determination of the quota system has the disadvantages of hysteresis and deviation,leading to deviations in the compiled expressway project estimates.The increasing investment scale and contracting mode also put forward higher requirements for the management and control of investment funds.In order to solve the above problems,through literature analysis and method research,the prediction model of expressway project estimate is established based on XGBoost algorithm model and cost indicator system,and their advantages are analyzed.Using a large number of screened indicators and project estimate information as the sample library,the XGBoost algorithm model is optimized by grid search and cross validation,the relevant parameter combination of the model is adjusted,and the parameter settings of the XGBoost algorithm model are optimized.Using 3 commonly used prediction accuracy evaluation indicators(RMSE,MAE,R2)as the evaluation system,the prediction accuracy,stability and deviation of the model are evaluated.Finally,a new cost prediction scheme is proposed,and the established XGBoost algorithm prediction model is sampled and verified to show the prediction ability of the algorithm model in the process of expressway project estimate prediction.It has a certain promotion effect on popularizing the cost indicator system and improving the accuracy of expressway project cost prediction.Meanwhile,a detailed analysis of the influencing factors of the cost indicators is conducted,the importance of characteristics as the indicators are analyzed and explored,i.e.,the important influencing factors of the expressway cost indicators in China are explored,the potential relationship between the influencing factors and the cost indicators are explored,which provides the investors and designers with new auxiliary reference indicators and research directions for estimate compilation.
作者 郭婧娟 刘曜玮 GUO Jing-juan;LIU Yao-wei(School of Economics and Management,Beijing Jiaotong University,Beijing 100044,China;CCCC Investment Co.,Ltd.,Beijing 100101,China)
出处 《公路交通科技》 CAS CSCD 北大核心 2023年第3期58-68,共11页 Journal of Highway and Transportation Research and Development
基金 住建部软科学研究项目(2021-R-089)。
关键词 道路工程 预测模型 XGBoost算法 工程概算 造价指标 road engineering prediction model XGBoost algorithm project estimate cost indicator
  • 相关文献

参考文献14

二级参考文献61

共引文献112

同被引文献51

引证文献5

二级引证文献5

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部