Rule Generation Method for Deriving Characteristics of Customers in Customer Support Center

Abstract

In the financial field, the use of customer information analysis systems is important to the customer support section to improve customer service. A customer information analysis system is composed of an analysis system and a storage system. It pays attention to data mining technology that analyses the customer confluence. But when the above technique was applied to actual data, it was found that in some cases, rules that included parity expressions were in the majority. In other cases, most rules were formed from the character of specific demographic clusters.
A rule generation support method was proposed for deriving useful rules from rules generated by the rule induction method. A prototype of the customer information analysis system was developed using the proposed method. This prototype was applied to actual data from the customer support section of the financial field, and confirmed its usefulness.