An Attrition Analysis of Hospital Customers and Development of Prediction Model on the Basis of Data mining

À±Çý¿ø1), Ãµº´Ã¶*2), ±è³²¼ö3)

1)°í·Á´ëÇб³ ÀÇ°ú´ëÇÐ º¸°Ç´ëÇпø
2)°í·Á´ëÇб³ ÀÇ°ú´ëÇÐ º¸°Ç´ëÇпø
3)¼­¿ï¾Æ»êº´¿ø °ÇÁø¿î¿µÆÀ

Abstract : The objectives of this research were, concentrating on the matter of increasing churns, especially competitor churn in the recent hospitals, to understand the characteristics of the customer group who are expected churn in case competing companies do aggressive sales promotion, to find the associated factors of their breakaway and to prepare practical marketing strategy to keep the existing customers. The data mining techniques we used were decision tree, neural network, and logistic regression analysis. Among these three techniques, the logistic regression model showed the best prediction power in ROC curve verification. This model showed about 92.1% accuracy and might be used as a basis for developing a preventive measure for the customer are at risk of churn.
keyword : data mining, logistic regression, decisi on tree, neural network, prediction power

¿¬Á¦ºÐ·ù

, ,


÷ºÎÇÑ ÆÄÀÏ

kosmi05_Hyewon Yoon.hwp

÷ºÎÇÑ e-posterÆÄÀÏ

Åõ°í³»¿ëÀ» »èÁ¦ÇϽ÷Á¸é "»èÁ¦"¹öÆ°À» ´©¸£½Ã°í ´Ù½Ã Á¢¼öÇÏ¿© Áֽñ⠹ٶø´Ï´Ù.

 ¾Ï È£

´ëÇÑÀÇ·áÁ¤º¸ÇÐȸ
Copyright (c) 2002 by The Korean Society of Medical Informatics