A Rolling-Window Framework for Churn Prediction and Behavioral Driver Identification
Researchers have developed a new framework for predicting customer churn using a rolling-window approach. This method models customer behavior within a 30-day observation window to estimate churn risk as activity evolves. The framework integrates both feature-based and sequence-based learning techniques, demonstrating robust performance on a real-world dataset with accuracy reaching 87.6% and ROC-AUC of 0.94 for the feature-based model. AI
IMPACT Provides a robust and interpretable method for predicting customer churn, potentially improving retention strategies in service environments.