基于PCA-ANN的水利工程标高金预测模型  

Prediction model for mark-up of water conservancy projects based on PCA-ANN

作　　者：石林林[1,2] 

机构地区：[1]河海大学商学院,江苏南京211100 [2]河海大学项目管理研究所,江苏南京211100

出　　处：《水利经济》2014年第3期52-55,66,共5页Journal of Economics of Water Resources

基　　金："十一五"国家科技支撑计划重大项目(2006BAB04A13);江苏省教育厅高校哲学社会科学基金(07SJD630006)

摘　　要：在水利工程投标竞争中,施工企业投标报价的合理性是赢得该项目的重要前提,而标高金的准确性直接影响投标报价的合理性。采用基于主成分分析的BP神经网络方法,建立了标高金预测模型。利用主成分分析(PCA)方法,对标高金的影响因素进行降维,在保留原始数据信息的条件下,通过消除样本数据间相关性、减少模型输入变量的数目,简化网络结。利用BP神经网络(ANN)寻找数据的内在联系和规律,客观而较为准确地预测出标高金。根据该模型,以某承包商历年来水利工程投标报价的原始数据为基础,对标高金进行了预测,结果表明模型预测结果满足实际需要。In competitive bidding of water conservancy projects,the rationality of construction companies 'offer is the prerequisite to a successful bidding. The accuracy of mark-up directly relates to the rationality of the offer. A prediction model for the mark-up is established by using the BP artificial neural network method based on the principal component analysis(PCA). The dimension of the factors to affect the mark-up is reduced by means of the PCA method,that is,under the conditions of retaining the original data,the network nodes are simplified by eliminating the correlation between the sample data and reducing the number of model input variables. The BP artificial neural network is adopted to search for the internal relation and law of the data so as to objectively and more accurately predict the mark-up. According to the proposed model, the mark-up of an example is predicted by using the original data of a contractor's offer of water conservancy projects for years. The results of the proposed model are practical.

关 键 词：水利工程 标高金 主成分 BP神经网络 

分 类 号：F284[经济管理—国民经济]