基于GS-SVR的架空输电线路工程投资估算预测研究

Investment estimation forecast of overhead transmission line project based on GS-SVR

传统的投资估算编制模式存在过度依赖定额的现象,随着大量工程造价数据的积累,利用其实现投资估算,以弥补传统定额计价模式的不足,能够对建设项目工程造价起到总体控制作用。文章以架空输电线路工程为例,基于支持向量回归机(Support Vector Regression,SVR)研究架空输电线路工程投资估算问题。结果表明:通过选取影响架空输电线路工程投资估算的主要指标,构建基于SVR的架空输电线路工程投资估算模型,并利用改进的网格搜索法(Grid Search,GS)优化模型参数,得到基于GS-SVR的投资估算预测模型;与传统的线性回归和SVR模型相比,GS-SVR模型表现出更为良好的性能。

The prevalent traditional mode of investment estimation exhibits an overreliance on fixed quotas.With a substantial accumulation of engineering cost data,addressing this issue by utilizing this data for investment estimation becomes a pressing need for construction entities,aiming to rectify the shortcomings of the traditional quota-based pricing model.Rational investment estimation can play a crucial role in overall control for construction projects.This paper,using the overhead transmission line project as an illustration,employs Support Vector Regression(SVR)to delve into the investment estimation predicament.The research involves selecting key indicators affecting the investment estimation of overhead transmission line project.Subsequently,a model for investment estimation is constructed based on SVR,and the improved Grid Search(GS)is employed to optimize the parameters of the investment estimate model,resulting in the GS-SVR investment estimation model.The results reveal that the GS-SVR model exhibits commendable performance compared to traditional linear regression and SVR models.