Under the sponsorship of the National Social Science Fund (Approval No. 17ZDA110), this project is led by Professor Pang Xun, focusing on theories and methods of geopolitical risk prediction as well as the development of an early warning system. Three main components of this project are listed as follows:

1) Classification of Geopolitical Risks: Based on their characteristics, causes, and dimensions of preventive and control measures, geopolitical risks can be grouped into five categories, including terrorism, international conflict risks, domestic security risks, domestic governance risks, and acquisition and dissemination of weapons of mass destruction.

2) Improvement of Geopolitical Risk Prediction Methods: Improving existing methods necessitates synthetizing qualitative and quantitative methods based on forecasting techniques commonly employed in the domain of global security and development. Specifically, this entails integrating qualitative information such as knowledge from experts into quantitative prediction models using Bayesian elicited priors. Drawing on relevant databases and data mining methods for quantitative information, the project combines machine learning tools with expert knowledge in variable selection, dimensionality reduction and model training for the purpose of constructing a geopolitical risk prediction model.

3) Validation and Refinement: The predicted outcomes of the model will be cross-validated with real-world outcomes. The prediction model will be further refined using case studies. As such, the project attempts to provide a conceptual framework as well as a comprehensive, target-specific and accurate early warning system and response mechanism for geopolitical risks in the "Belt and Road" region.