Agent-Based Decision-Support Framework for Credit Risk Early Warning

Credit risk early warning is a complicated process. Early studies of prediction decision support systems (DSS) offer a passive form of support, where users needed to have full knowledge on how to use the data sources and other tools to perform all necessary operations. Agents are considered have capabilities of observe the environment and learn from previous experience to make their own decision, and could building active cooperative intelligent systems. This paper proposes a framework of a multi-agent decision support system (MADSS) used in credit risk prediction. The structure of a single agent is defined, and the mechanism of decision making process is designed. The framework supports model training, risk prediction and results analyzing based on financial information of customer. We outline the design principles and develop architecture for MADSS