Analytical CRM in banking and finance using SVM: a modified active learning-based rule extraction approach
by M.A.H. Farquad; V. Ravi; S. Bapi Raju
International Journal of Electronic Customer Relationship Management (IJECRM), Vol. 6, No. 1, 2012

Abstract: This paper presents advancement to modified active learning-based approach in an eclectic framework for extracting if-then rules from support vector machine (SVM) for customer relationship management (CRM) purposes. The proposed approach comprises of three major phases: 1) feature selection using SVM-RFE (recursive feature elimination); 2) active learning for synthetic data generation; 3) rule generation using decision tree (DT) and Naive Bayes tree (NBTree). Finance problems solved in this study are churn prediction in bank credit cards customers and fraud detection in insurance. Based on sensitivity measure, the empirical results suggest that the proposed modified active learning-based rule extraction approach yielded best sensitivity and length and number of rules is reduced resulting in improved comprehensibility. Feature selection leads to the most important attributes of the customers and extracted rules serves as early warning system to the management to enforce better CRM practices and detect/avoid possible frauds.

Online publication date: Sat, 16-Aug-2014

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Electronic Customer Relationship Management (IJECRM):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com