Bayesian Analytics for Estimating Risk Probability in PPP Waste-to-Energy Projects

Abstract

Appropriate risk analysis and management is critical to the overall success in public–private partnership (PPP) projects, in which one of the key issues lies in an accurate estimation of the risk occurrence probability. Traditionally, this probability is estimated either relying on experts’ judgments or historical data. The estimation may not be accurate due to the subjective nature of the former and the data sparsity of the latter. In this research, a Bayesian analytic approach is taken to forecast risk occurrence probability, combining experts’ judgments and historical data. This Bayesian approach consists of four main steps: (1) data collection, (2) modeling prior probability, (3) modeling posterior probability, and (4) multiupdating and analytics. This approach can achieve a more accurate estimation of risk occurrence probability compared with only relying on experts’ judgments or historical data because the subjectivity of experts’ judgments is mitigated by incorporating observed real data, and the data sparsity is supplemented by experts’ judgments. This model is applied to forecast the probability of several critical risks in PPP waste-to-energy (WTE) incineration projects in China, and the results demonstrate its feasibility and applicability for targeted solutions in risk response and allocation.