Author(s)

Kesen Zhang

Corresponding Author

Kesen Zhang

ABSTRACT

A life insurance policy can pay off any debt one left behind, which will put a burden on the family. Debts such as mortgages, credit cards, car loans, and even the funeral expenses can have a significant impact on one's family and their lifestyle. In a one-click shopping world with on-demand everything, only 40% of US household own individual life insurance due to the antiquated application process from life insurance companies. Customers need to provide extensive information for identification of risk classification and eligibility, which includes scheduling medical exams, a process that takes an average of 30 days. Gradually, customers tend to fall off. Here we propose an approach named ‘XGBoost Based on Offset' estimator to solve such a problem. XBO is a predictive model that accurately classifies customer's level using a more automated approach. With technology frameworks and the application of mathematical statistics, we show that XBO estimator makes the identification quicker and less labour-intensive for new and existing customers while maintaining privacy boundaries. By automatically separating clients to eight ranks, this model can facilitate the pricing process of life insurance policy and can make a difference for the life insurance companies' risk management.

KEYWORDS

Customer Assessment, XBO model, Life Insurance Company