APPLYING ECONOMIC MEASURES TO LAPSE RISK MANAGEMENT WITH MACHINE LEARNING APPROACHES

Abstract Modeling policyholders’ lapse behaviors is important to a life insurer, since lapses affect pricing, reserving, profitability, liquidity, risk management, and the solvency of insurer. In this paper, we apply two machine learning methods modeling. Then, evaluate performance these along with popular statistical by means accuracy profitability measure. Moreover, adopt an innovative point view on prediction problem that comes from churn management. We transform classification into regression question then perform optimization, which new aforementioned four large real-world insurance dataset. The results show Extreme Gradient Boosting (XGBoost) support vector outperform logistic (LR) tree respect statistic accuracy, while LR performs as well XGBoost in terms retention gains. This highlights importance proper validation metric when comparing different methods. optimization after transformation brings out significant consistent increases economic Therefore, insurer should conduct its objective achieve optimal