论文部分内容阅读
采用基于结构风险最小化思想的支持向量回归(SVR)模型预测了灌注桩的中标价格。在分析灌注桩的成本动因的基础上,通过历史项目清单预算确定了17个回归自变量,根据历史项目中标价格(因变量)和成本动因(自变量)来估算当前项目价格。利用该模型对福建省某高速公路灌注桩进了行价格估算,并与基于定额的清单预算价格以及多元线性回归估算价格进行了对比。结果表明:估算结果相对于其他两种方法明显接近实际中标价格,采用SVR模型可以大大提高估算的准确性,有效控制造价,提升公路工程造价管理水平。
The support vector regression (SVR) model based on structural risk minimization was used to predict the bidding price of bored piles. Based on the analysis of the cost driver of the bored pile, 17 regression independent variables were determined according to the list of historical items, and the current project price was estimated according to the bid price (dependent variable) and the cost driver (independent variable) of the historical project. This model is used to estimate the price of bored piles in an expressway in Fujian Province, and compared with the quoted list price and multivariate linear regression estimated prices. The results show that compared with the other two methods, the estimation result is obviously close to the actual bid price. Using SVR model can greatly improve the accuracy of estimation, effectively control the cost and improve the management level of highway project cost.