Early Warning Systems (EWS) are essential in the dynamic information management environments of banks and financial institutions to monitor credit portfolios and detect early indicators of financial distress. The literature shows that EWS typically focus on identifying negative indicators, lacking research on positive metrics. Banks and financial institutions can enhance their operations by utilising artificial intelligence to incorporate automated Customer Segmentation (CS) techniques to identify positive indicators, such as up-selling opportunities and enhance information systems management. This study explores how integrating EWS into a new CS model can reveal possible business opportunities in commercial credit portfolios, with an emphasis on segment orientation, identifiability, and actionability. We utilise a dataset from a major European commercial bank to experiment our approach for combining EWS with CS to provide information management solutions. We utilised K-Means and DBSCAN clustering techniques with EWS indicators, in addition to Principal Component Analysis (PCA) for dimensionality reduction, to efficiently categorise the client base. We conducted risk–reward and risk exposure evaluations and analyses to improve clients’ understanding of health. The assessment of clustering outcomes demonstrated that PCA is superior in improving the compactness and distinctiveness of clusters. Our research provides valuable insights into strategic risk monitoring by integrating EWS and CS. This allows banks and financial institutions to customise their services for various client groups, enhancing strategic decision-making and potentially boosting profitability.
Amato et al. (Fri,) studied this question.