摘要
为选择、筛查、验证和调整价差系数,以用于计算价差预备费,本文结合常用建设项目工程造价费用划分树结构,引入单位根检验和脉冲响应函数,建立过滤机制模型。基于模型筛选、匹配相应系数,对各系数形成的时间序列进行平稳性检验和均值回归能力检验。研究发现,该过滤模型具有良好的筛查能力和多维检验效果,实证分析发现,当前用于系数调差的绝大多数时间序列需进行差分处理后方可使用。尤其,当存在时间序列长度不符合预测技术基本要求时,需进行样本容量扩充和数据再检验。与传统的将系数与基础价格直接整合的调差方法相比,本文提出了一种适用于建设工程尤其是境外铁路工程系数调差的过滤机制模型,该过滤模型避免了因忽视统计学原理而引起的变量选取的非稳健性,减少了预测的随意性,降低预测结果的失真。
For selecting,screening,validating and adjusting indices for calculation of escalation,combing with the traditional cost classification tree of construction projects,this article introduces the unit root tests and impulse response functions to establish a filtering mechanism model.The selection and allocation of corresponding indices,as well as the validation of stationary test and mean-reverting behavior test of time series,are based on the filtering mechanism mode l.The filtering mechanism proposed in this article reveals good screening ability and multi dimension test effects.The empirical analysis shows that most of the time series used for the price difference adjustment should be first-differenced or second-differenced.In particular,when the length of the time necessary for a shock to cost series to dissipate does not meet the basic requirements of prediction,it is essential to expand the sample size and re-test.This article puts forward a filtering mechanism model which can be used to adjust the price differences,especially for overseas railway engineering.T h e model avoids the non-stationary when adopting the variables due to the neglect of the basic principles of statistics,reduces the arbitrariness of prediction and avoids the distortion of prediction results,comparing with integrating the indices with capex directly.
作者
徐新
XU Xin(China Railway Siyuan Survey and Design Group Co Ltd,Wuhan 430063,China)
出处
《土木工程与管理学报》
北大核心
2018年第2期148-154,169,共8页
Journal of Civil Engineering and Management
基金
中铁第四勘察设计院集团有限公司科研课题(2011K18)
关键词
铁路工程
平稳性
价差系数
脉冲响应
价差预备费
railway engineering
stationary
indices for price difference
impulse response
price difference reserve fee
