摘要
由于公路工程造价估算具有多因素、高维非线性、随机性等特点,传统的造价估算方法无法快速得到准确结果,为此提出基于支持向量机的公路工程造价快速估算模型。首先结合公路工程特性,确定造价的主要影响因子,然后通过LibSVM工具箱建立造价与主要影响因子之间的非线性映射关系,同时采用交叉验证优化支持向量机参数。最后进行实证检验,并与BP神经网络估算方法进行对比分析,结果表明支持向量机提高了估算精度,可以更好地应用于公路工程造价估算。
Because the cost estimation of highway engineering has the characteristics of multi factors,high-dimensional nonlinearity and randomness,the traditional cost estimation method can not get accurate results quickly.Therefore,a fast estimation model of highway engineering cost based on support vector machine is proposed.Firstly,combined with the characteristics of highway engineering,the main influencing factors of cost are determined.Then,the nonlinear mapping relationship between cost and main impact factor is established through LibSVM toolbox,and the support vector machine parameters are optimized by cross-validation.Finally,the empirical test is carried out and compared with the BP neural network estimation method.The results show that the support vector machine improves the estimation accuracy and can be better applied to highway engineering cost estimation.
作者
邹鑫
李远富
杨昌睿
樊惠惠
吴文芊
ZOU Xin;LI Yuanfu;YANG Changrui;FAN Huihui;WU Wenqian(School of Civil Engineering,Southwest Jiaotong University,Chengdu 610031,China;Key Laboratory of High Speed Railway Line Engineering,Ministry of Education,Chengdu 610031,China)
出处
《交通科技》
2019年第6期19-21,共3页
Transportation Science & Technology
关键词
公路工程
造价估算
交叉验证
支持向量机
BP神经网络
highway engineering
cost estimation
cross-validation
support vector machine
BP neural network