基于BP和SVM变量筛选的LC技术特征对项目绩效比较研究  

Research on GA- BP with SVM for Coupling of Lean Construction Technology and Project Performance

作　　者：李书全[1] 胡少培 胡本哲 

机构地区：[1]天津财经大学商学院,天津300222

出　　处：《科技管理研究》2015年第7期174-179,共6页Science and Technology Management Research

基　　金：国家自然科学基金项目"精益建设(LC)技术采纳行为与决策模型研究"(71171140);天津市科委科技发展战略研究计划项目"建筑施工企业安全投资决策模型研究及应用"(12JCZDJC34900);天津财经大学重点项目"基于社会资本的建筑安全预警研究"(ZD1306)

摘　　要：为探究精益建设技术与项目绩效之间的内在作用机理,构建基于BP和SVM变量筛选的6S、可视化管理、最后计划者等7种精益建设技术与知识能力、财务、业主等5个项目绩效分项指标和综合指标的耦合模型。仿真结果表明:在精益建设技术特征与项目绩效分项指标的耦合模型仿真分析中,基于GA-BP的预测模型比标准BP神经网络模型精度要高;在精益建设技术特征与项目绩效综合指标的耦合模型仿真分析中,基于SVM的预测模型比GA-BP的预测模型精度要高。另外,利用BP和SVM结合MIV算法进一步探究不同精益建设技术对项目绩效各指标和综合指标的影响程度。研究结果为项目利益相关者提高项目管理绩效提供决策支持。In order to explore the internal mechanism between inquiry lean construction technology and project perform-ance,the paper constructs 7 kinds of lean construction technology such as the screening of BP and SVMvariables 6S,visu-al management,the last planner,and the coupling model consisting of comprehensive indexes and 5 project performance in-dexes such as knowledge ability,finance,owners.The simulation results show that：in the analysis of coupled model simu-lation technology characteristics of lean construction and project performance sub index,prediction model of GA -BP is higher than standard BP neural network model based on the analysis of precision;in the analysis of coupled model simula-tion comprehensive indicators of performance technology characteristics of lean construction and project,the predictive mod-el of SVMis higher than the precision of forecasting model GA -BP based on.In addition,using the BP and SVM com-bined with MIV algorithm,the paper further explores the influence of different lean construction techniques on each index and comprehensive index of project performance.Research results provide decision support for the project stakeholders to improve the performance of project management.

关 键 词：精益建设技术特征 项目绩效 遗传神经网络(GA—BP) 支持向量机(SVM) 变量筛选 

分 类 号：F426.92[经济管理—产业经济] F274