Data Mining Model to Analyse Crm in Banking Sector

Abstract

The Banking business is profoundly serious, dynamic and subject to quick change. Thus, Banks
are being pushed to comprehend and rapidly react to the individual needs and needs of their
clients. Client relationship the executives (CRM) is the general procedure of abusing client
related data and utilizing it to improve the income stream from a current client. Information
mining methods are utilized to remove significant client data from accessible databases. 

The goal of the research is to propose an efficient Customer Relationship Management data
mining model for the prediction of customer relationship in the domain of banking applications,
specifically to profile loyal and lost bank customers based on their RFM (Recency, Frequency
and Monetary) score that help to address the problem domain. The subject of this research is
Commercial Bank of Ethiopia customer transaction and demographic information of more than
70,000 customers. Within the built model, different segments of customers are studied and
evaluated based on their common attributes since customer grouping is the main part of customer
relationship management. Customer classification and grouping is done by combining the
Support Vector Machine (SVM) ensemble classifier Random Forest and K-means methods along
with the application of RFM (recency, frequency, and monetary) model. The prediction of
customers is also done using logistic regression model. The outcomes from this examination
were empowering, which fortified the conviction that applying information mining strategies
could undoubtedly bolster Customer Relationship Management exercises at Commercial Bank of
Ethiopia. Later on, more division contemplates utilizing segment and other client's data and
utilizing other grouping and arrangement calculations could yield better outcomes.