한국경영과학회:학술대회논문집 (Proceedings of the Korean Operations and Management Science Society Conference)
- 대한산업공학회/한국경영과학회 2002년도 춘계공동학술대회
- /
- Pages.306-311
- /
- 2002
A Personaliz Customer Retention Procedure For Internet Game Site Based on the Self-Organizing Map and Association Rule Mining.
- Song Hee Seok (Graduate School of Management, Korea Advanced Institute of Science) ;
-
Kim Jae Kyeong
(School of Business Administration, KyungHee University) ;
- Kim Soung Hie (Graduate School of Management, Korea Advanced Institute of Science) ;
- Chae Kyung Hee (School of Business Administration, KyungHee University)
- 발행 : 2002.05.01
초록
This paper propose a personalized defection detection and prevention procedure based on the observation that potential defectors have tendency to take a couple of months or weeks. For this purpose, possible states of customer behavior are determined from past behavior data using SOM (Self-Organizing Map). For the evaluation of the proposed procedure, a case study has been conducted for a Korean online game site. The result demonstestes that the proposed procedure can assist defection prevention effectively and detect potential defectors without deterioration of prediction accuracy comparison to prediction by MLP. Our procedure can be applied to various service industries that can capture fluent customer behavior data such as telecommunications, internet access services, and content services, too.
키워드