A general framework for the identification and categorization of risks: an application to the context of financial markets

Micha Bender and Sven Panz

Abstract

Risk represents a significant part of human interaction and must be considered in decision-making processes across diverse business and research areas. Further, the disregard or unawareness of certain risks may result in inappropriate decision-making processes and inadequate risk management practices that may negatively influence firms’ performance. This paper is, to the best of our knowledge, the first to develop an algorithm-based and generally applicable framework that generates an extensive and integrated identification and categorization scheme of certain risks by using text mining and machine learning approaches. To demonstrate the applicability of our framework, we apply our approach to the context of financial markets, identify 193 financial market risks and sort them into five categories by using common machine learning techniques. To evaluate the general applicability, we additionally apply our derived framework to the context of information systems. Finally, we obtain strong indications of the robustness and superiority of our derived framework by benchmarking it against more manual risk identification techniques and other clustering approaches.