Reconciling performance and interpretability in customer churn prediction using ensemble learning based on generalized additive models.
Koen W. De BockDirk Van den PoelPublished in: Expert Syst. Appl. (2012)
Keyphrases
- ensemble learning
- customer churn
- churn prediction
- prediction accuracy
- ensemble methods
- random forest
- prediction model
- generalized additive models
- generalization ability
- base classifiers
- customer relationship management
- decision trees
- unlabeled data
- regression model
- pattern recognition
- concept drift
- rule induction
- bp neural network
- learning algorithm
- feature set
- active learning
- credit risk
- training set
- decision making
- feature selection