摘要
基于建设项目动态联盟候选投标项目评价的内涵分析,确定了候选投标项目评价的决定因素,构建了候选投标项目评价的指标体系。首先通过计算欧氏贴近度,剔除了贴近度较小的指标,然后采用主成分分析将众多指标进行综合,消除样本间的信息重叠,降低RBF网络的输入维数。针对候选投标项目评价系统的非线性特征,采用RBF网络高度非线性映射能力,对某建设项目动态联盟的候选投标项目进行了评价。评价结果表明EN、PCA与RBF网络相结合的方法比PCA与RBF网络相结合的方法及单纯的RBF网络方法具有较高的精确度和较好的拟合效果。整个数据处理过程用软件完成,成本低廉、运算速度快捷,能够克服数据处理流程的复杂性,具有较好的实用性。
Based on the analysis on the meaning of evaluating candidate bidding projects of virtual enterprises of construction project,we identify the determinants of evaluating candidate bidding projects,and establish the evaluation criteria system for candidate bidding projects.Firstly,through computing Euclid nearness,we eliminate two criteria whose Euclid nearness is too small.Then,through principal component analysis,we synthesize numerous criteria,eliminate information overlapping of the sample,and reduce the input dimension of RBF network.According to the nonlinear feature of candidate bidding projects evaluation system,using RBF network altitudinal nonlinear map,we evaluate candidate bidding projects for a virtual enterprise of construction project.The results show that the conjoint method-EN,PCA and RBF network is more precise and fits better than PCA and RBF network method and the single RBF network method.The data is computed by software which can make the whole computing process cheap and swift,and can overcome the complexity of the data computing process,which has good practicability.
出处
《管理评论》
CSSCI
北大核心
2010年第2期121-128,共8页
Management Review
基金
国家自然科学基金项目(70571038)
关键词
候选投标项目评价
建设项目动态联盟
欧氏贴近度
主成分分析
RBF神经网络
candidate bidding projects evaluation
virtual enterprise of construction project
Euclid nearness（EN）
principal component analysis（PCA）
RBF neural network
