承包商预测业主支付违约行为的多元线性回归模型构建

发布时间：2018-02-09 08:14

本文关键词： 承包商业主支付违约行为影响因素多元线性回归模型逐步回归法　出处：《昆明理工大学》2014年硕士论文　论文类型：学位论文

【摘要】：业主支付违约行为是中国建设领域多年来面临的难题,已严重影响了建设市场的正常秩序,阻碍了承包商的健康发展,破坏了经济稳定和社会和谐。因此,寻找业主支付违约行为的主要影响因素,使承包商在投标决策时能快速预测业主支付违约行为具有重要的现实意义。文章站在承包商的角度,在其自身实力满足业主投标要求的前提下分析业主支付违约行为。文章围绕业主支付违约行为这一主题探讨了业主支付违约行为的影响因素及其分类体系,影响业主支付违约行为的主要微观因素,构建业主支付违约行为的预测模型三个研究问题。首先,文章通过文献回顾和梳理,总结出较全面的业主支付违约行为影响因素并对其进行了两分法的分类,得到宏观因素3个,微观因素11个。接着,在影响因素分类的基础上,通过专家打分法确定了5个主要微观影响因素分别是业主融资能力和本身的资金实力、业主的声誉、双方信任程度、合同签订情况、承包商的履约表现。最后,文章将5个主要微观影响因素作为多元线性回归模型的自变量建立理论框架和研究假设,并以此为基础设计问卷,通过统计软件SPSS17.0判断问卷的信度与效度和分析数据,采用逐步回归法估计出多个模型,根据模型拟合数据程度以及总体回归方程和回归系数的显著性,选择出最优模型,并且依次检验模型的基本假定,诊断数据中是否存在异常值,用建立好的模型预测业主支付违约行为。研究结论显示,模型中业主融资能力和本身的资金实力的回归系数为-0.471,承包商的履约表现的回归系数为-0.262,业主的声誉的回归系数为-0.257,双方信任程度的回归系数为-0.206,合同签订情况的回归系数为-0.120,常数项为7.398。文章在运用多元线性回归法构建预测模型研究业主支付违约行为的问题上具有创新性,有助于承包商在投标时做出理性决策。
[Abstract]:The breach of payment by the owner has been a difficult problem in the construction field of China for many years, which has seriously affected the normal order of the construction market, hindered the healthy development of the contractor, destroyed the economic stability and social harmony. It is of great practical significance to find out the main influencing factors of the employer's payment breach of contract, so that the contractor can predict the employer's payment breach of contract quickly when making the bidding decision. The article stands at the contractor's point of view, On the premise that its own strength meets the requirements of owner's bidding, this paper analyzes the owner's payment breach of contract. This paper discusses the influencing factors and its classification system of the owner's payment breach of contract, focusing on the subject of the owner's payment breach of contract. The main microcosmic factors that affect the owner's payment breach of contract and the construction of the forecasting model of the owner's payment default behavior are studied. First of all, the article reviews and combs through literature review, This paper summarizes the influencing factors of the owner's payment breach of contract and classifies them by dichotomy, and obtains 3 macro factors and 11 micro factors. Then, on the basis of the classification of influencing factors, Through the expert scoring method, it is determined that the five main microcosmic influencing factors are the owner's financing ability and his own capital strength, the owner's reputation, the degree of mutual trust, the signing of the contract, the performance of the contractor. Finally, In this paper, five main microcosmic factors are used as independent variables of multivariate linear regression model to establish theoretical framework and research hypotheses. Based on this, a questionnaire is designed to determine the reliability, validity and analysis data of the questionnaire by statistical software SPSS17.0. Several models were estimated by stepwise regression method. The optimal model was selected according to the degree of fitting data and the significance of the total regression equation and regression coefficient, and the basic assumptions of the model were tested in turn. If there are outliers in the diagnostic data, the established model is used to predict the defaulting behavior of the owners. In the model, the regression coefficient of the owner's financing ability and his own capital strength is -0.471. the regression coefficient of the contractor's performance is -0.262, the regression coefficient of the owner's reputation is -0.257, the regression coefficient of mutual trust is -0.206, and the contract signing condition. The regression coefficient is -0.120, and the constant term is 7.398.This paper is innovative in using the multivariate linear regression method to build a prediction model to study the defaulting behavior of the owner. It helps the contractor to make rational decision in bidding.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F426.92;F224

【参考文献】