Customer Churn Prediction Using Local Linear Model Tree for Iranian Telecommunication Companies

For winning in global competition، companies need to recognition and monitoring of customer''s behavior to forecast their behavior and desires earlier than competitors. This research tries to recognize the attributes which lead to customer churn. For this، behavior of 3150 subscribers of an Iranian mobile operator، has observed during one year and trends of them has analyzed by a customized LLNF model. For this purpose، the application of the locally linear model tree (LOLIMOT) algorithm، which integrates the advantage of neural networks، tree model and fuzzy modeling، was experimented. Results suggest that dissatisfaction of customer، his/her usage from services and demographic attributes have significant effect on decision to churn or retention. Furthermore، the active or inactive subscriber situation has mediation effect on his/her retention.