Prediction Model Of Private Capital's Willingness To Participate In PPP Projects And Analysis Of Its Influencing Mechanism Based On Machine Learnin

The PPP(Public-Private Partnership)model is a relatively new financing model that has gained rapid popularity in China.Unlike the PPP models used in other countries,the PPP model in China focuses on the collaboration between social capital and local governments.including the state-owned capital-dominated public economy and the non-public economy dominated by private capital.Private capital,as an important indicator of measuring the vitality of China’s market economy,has significant importance in the promotion of the PPP model.On the one hand,the participation of private capital can alleviate the financial pressure and transfer risks of local governments.Private capital has access to more efficient management models and operational practices,which can significantly enhance the quality of public services.In reality,however,the participation of private capital has been consistently low,with a large number of PPP projects still being dominated by state-owned enterprises.Furthermore,regardless of the scale or quantity of private enterprise participation in PPP projects,the participation rate has been declining.Therefore,promoting the investment willingness of private capital in PPP projects has become an important issue for the benign development of the PPP model.This study starts from the idea of big datadriven project management,establishes a predictive model of private capital participation in PPP project investment willingness,this involves a comprehensive analysis of the factors that influence the participation of private capital in PPP(Public-Private Partnership)projects.and explores its influencing mechanism,providing decision-making reference for the government to stimulate private capital investment willingness.This study first used literature research and text analysis to determine the feature set of private capital participation in PPP project investment willingness from four aspects: project attributes,local governments,macroeconomics,and market environment.Then,the data in the feature set was obtained through network crawlers and manual collection to form a data set of private capital participation in PPP project investment willingness.After preprocessing the data,it was found that the data set had problems such as high feature dimensionality,imbalanced data distribution,and significant data noise.Therefore,the Dragonfly Algorithm was selected for feature selection engineering,the improved KSMOTE algorithm was used to deal with imbalanced data sets,and the Stacking mixed integration strategy was used to improve the integration algorithm to enhance the model’s generalization ability and reduce data noise.Next,the RF,GBDT,ADABoost,Light GBM,and XGBoost algorithms were used for integration to obtain four mixed integration models,RF-XGBoost、ADABoost-XGBoost、Light GBM-XGBoost and GBDT-XGBoost,and establish a predictive model of private capital participation in PPP project investment willingness.The model was evaluated using indicators such as ACC,AOC,and F1-score values.It was found that the mixed integration models,after feature selection and imbalanced data processing,had good predictive performance,with the Light GBM-XGBoost model having the best predictive effect.After that,the SHAP model was used to perform a visual interpretability analysis of the Light GBM-XGBoost model,exploring how various features affect private capital participation from the perspective of features and samples.The results showed that the main features affecting private capital participation willingness were the government’s experience in participating in PPP projects,in addition to project risk present value,regional economic level,corruption level,and the level of development of non-public economy variables also significantly affected private capital participation willingness.Finally,based on the research results,targeted suggestions were provided on how to enhance private capital participation in PPP project investment willingness,providing decision-making reference for the government to stimulate private capital investment willingness.