An effective hybrid learning system for telecommunication churn prediction

Highlights：

• A hybrid model learning system for telecom customer churn prediction is proposed.

• The hybrid model integrates weighted k-means clustering and a rule induction method.

• The effectiveness of the proposed system is estimated using ROC, AUC and accuracy.

• The proposed system achieves superior prediction performance against other models.

• The proposed system can be widely applied in applications of different realms.