基于贝叶斯网络的PPP项目风险预测
The Risk Prediction of the PPP Project in China Based on the Bayesian Network
孙慧 1王煊昂1
作者信息
- 1. 天津大学管理与经济学部,天津 300072
- 折叠
摘要
基于影响图的基本原理和PPP项目风险的相关研究,识别了 PPP项目风险传递链.然后使用142份调查问卷数据,运用期望最大化(EM)算法和累积分布函数(CDF)算法,构建并验证了 PPP项目风险贝叶斯网络模型.结果显示,该模型对投资收益、项目功能、溢出效应3个项目目标预测的准确性分别为77.81％、73.40％和79.15％.运用其学习、推理功能,推导出PPP项目的关键风险源的类型是项目因素和政府因素,影响投资收益的关键风险事件是使用量不足风险和施工成本超支风险.从某水务PPP项目的案例分析来看,模型在PPP项目投资及风险管理中具备一定应用价值.
Abstract
Based on the influence diagram and the related research on PPP project risk,the PPP project risk transmission re-lationship is identified.Then,142 questionnaires were used to obtain the data,and the expectation maximization(EM)algo-rithm and cumulative distribution function(CDF)algorithm were used to construct and verify the PPP project risk Bayesian network model.The results show that the prediction accuracy of this model is 77.81％,73.40％and 79.15％for invest-ment return,project function,and spillover effect,respectively.Using its learning and reasoning function,it is deduced that the key risk sources of PPP projects are project and government aspects,and the key risk events affecting investment returns are insufficient use and construction cost overrun risk.From the case analysis of a water PPP project,this model provides value for PPP project investment and risk management.
关键词
贝叶斯网络/PPP项目/风险预测Key words
Bayesian network/PPP project/risk prediction引用本文复制引用
出版年
2024
NETL
NSTL
万方数据