Abstrakt
In our ongoing project focusing primarily on bank clients, we suggest an innovative strategy that will overcome persisting shortcomings of the state-of-the-art methods. We propose a basic concept and principles of a novel solution inspired by both of the mentioned directions and based on the idea of analyzing a social network of clients.
We do not assume the existence of such a network (as this is a strong, often unrealistic assumption), but from a given set of client financial activities we infer their social network in order to analyze mutual client relationships and behavior. For this purpose, not just the traditional direct facts are incorporated, but also relationships inferred using similarity measures and statistical approaches, with possibly limited measures of reliability and validity in time.
Such networks enable analyses of client characteristics from a new perspective and can provide otherwise impossible insights.