摘要
为探究社会资本参与PPP意愿及特征,提出一个基于集成学习的可解释框架,通过SHAP模型解释预测结果,以增加模型透明度和可靠性。结果表明:基于SHAP可解释框架具有良好的预测和解释能力,有助于政府准确快速识别社会资本参与意愿以及特征;社会资本对于PPP项目所属领域具有选择偏好,更偏爱市政工程类项目;政府付费、宏观经济环境稳定以及开放的市场等更容易吸引社会资本的参与;过度超前基建投资反而会增加对社会资本挤出效应。据此,提出对策建议:建立与项目产出绩效相挂钩的回报机制;不能过度超前投资建设基础设施,以免造成对社会资本的挤出效应;处理好政府与市场之间的关系,营造良好营商环境,平等合作。
This paper proposes an interpretable framework based on ensemble learning to explore the willingness and charac-teristics of social capital to participate in PPP.The prediction results were explained through the SHAP model to increase its transparency and reliability.The results showed that,first,the SHAP interpretable framework has good prediction and inter-pretation capabilities,which could help the government accurately identify social capital's willingness and characteristics to participate.Second,the social capital prefers municipal engineering PPP projects.Third,government payments,a stable macroeconomic environment,and an open market were more likely key factors to attract the participation of social capital.Fourth,excessive investment in infrastructure would increase the crowding-out effect on social capital.Based on this,three countermeasures and suggestions were put forward.A return mechanism linked to project output performance should be es-tablished to attract private capital to participate in PPP projects.Investment in infrastructure should not be excessively ad-vanced to avoid crowding-out effects on social capital.A good business environment,equal partnership,and government role were conducive to promoting social capital's participation in PPP.
作者
沈俊鑫
李晓颖
张云华
SHEN Jun-xin;LI Xiao-ying;ZHANG Yun-hua(Faculty of Management and Economics,Kunming University of Science and Technology,Kunming 650093;Faculty of Civil Engineering and Mechanics,Kunming University of Science and Technology,Kunming 650093)
出处
《软科学》
CSSCI
北大核心
2023年第12期136-144,共9页
Soft Science
基金
国家自然科学基金项目(71964018)
云南产业发展研究项目(2022Z06)
云南省省院省校合作项目(SYSX201911)。
