발간년도 : [2014]
| 논문정보 |
|
| 논문명(한글) |
[Vol.9, No.3] A Hybrid Data Mining Classifier for Prediction Churn Customers |
|
| 논문투고자 |
Boo-Sik Kang |
|
| 논문내용 |
The activities for customer relationship retention are most essential process of customer relationship management. If we improve the rate of churn customers about 5%, it is known that the profit rate of the company raises about 100%. Prediction activities must be preceded for management of the churn customers. If we can predict churn customers in advance, we can minimize the loss because of
removing the churn factor. Data sets about churn customers have an imbalanced characteristic that the number of the churn customers is less remarkably than the number of the retention customers. One of the main issues on the models to predict churn customers is to increase prediction performance. This research deals with decision tree, neural networks, and SVM models those are mainly used to predict the churn customers and known as good prediction performance, and proposes a hybrid weighted data mining classifier. To test the prediction performance of proposed method, this study used 'churn data sets' in UCI Machine Learning Repository. Experimental results showed that the method was effective for prediction of churn customers. |
|
| 첨부논문 |
|
|
|
|
|
자료실 > 논문지 
