Stopping Rule을 이용한 자동차보험 위험집단 구분 및 적정 손해 수준 추정

초록

For the insurance pricing, variable selection which has different risk pattern and grouping for risk estimation should be considered first. Since a variable, such as gender, is an obvious classification measure, it is not appropriate for considering one of grouping criteria. However, an age variable has a wide range and the criteria for its grouping is ambiguous. Considering each age for risk estimation makes a credibility issue due to the number of customer in each cell. Moreover, the age variable has its unique characteristics that is depending on time and the order for the grouping should be reflected. In this research, the most effective way for the age variable grouping is proposed by considering the variable characteristics. More precisely, various grouping methods currently applied and new method 'The Stopping Rule' will be introduced. Using real insurance data, analysis results are given to compare the performance and also be shown that the properness and effectiveness of the proposed grouping method, ‘The Stopping Rule’.

키워드