An OWA-Powered Dynamic Customer Churn Modeling in the Banking Industry Based on Customer Behavioral Vectors

In recent years, the issue of churn has become of greater significance. This term refers to the customer abandoning the company. The management of customer relationships must take into account the dynamic and changing nature of client behavior over time. The models offered to identify and predict customer turnover can be categorized as either dynamic or static. Static models examine client behavior at a single moment in time, whereas dynamic models examine customer behavior across time. In order to model customer churn dynamically in the banking industry, we used bank customer data to analyze the dynamics of customer churn behavior, leveraging the power of information fusion theory methodologies, especially the ordered weighted averaging (OWA) operator. By extracting better-period customer behavioral vectors in different time intervals and gaining the most common behavioral vectors, we compute their similarity to the selected customers' behavioral vectors. In addition to identifying the most common behavioral vectors, especially the churned customers, we can measure the similarity between each selected customer outside the database and the most common behavioral patterns. Thus, measuring the churn value for each individual customer will become possible. The banking industry will gain a more dynamic customer relationship management system by analyzing the patterns of customer behavior changes.