Profit-driven weighted classifier with interpretable ability for customer churn prediction

Customer churn prediction methods aim to identify customers with the highest probability of attrition, improve the effectiveness of customer retention campaigns, and maximize profits. However, previous studies have relied on a single classifier, leading to suboptimal predictive results. To address this issue, we propose a novel profit-driven weighted classifier that integrates a weighted strategy with multiple profit-driven ensemble members. We employ an artificial hummingbird optimization algorithm to determine the optimal weight coefficients of the profit-driven ensemble members based on the expected maximum profit criterion. We then calculate the Shapley additive explanation value to further improve the interpretability of the proposed weighted classifier. We conducted experiments and statistical tests on eight real-world datasets from different industries. The results show that the proposed weighted classifier significantly improves profits compared with comparative classifiers and provides strong interpretability based on the Shapley additive explanation value.