基于神经网络的人身险公司经济资本计算方法探索  

Exploration of Neural Network Based Economic Capital Calculation Method for Life Insurance Companies

作　　者：胡安睿 HU Anrui(School of Finance,Nankai University,Tianjin 300350,China)

机构地区：[1]南开大学金融学院,天津300350

出　　处：《科技和产业》2023年第22期193-198,共6页Science Technology and Industry

摘　　要：“偿二代”二期工程的实施对保险公司风险管理提出了更高要求,经济资本的计算受到保险公司重视。嵌套随机模拟法是传统的经济资本计算方法,但是绝大多数公司无法承担该方法所需的时间成本和算力资源。本文提出一种基于神经网络的经济资本计算方法,并将其应用于最低年金给付保证的变额年金产品中。结果表明,相较于传统的嵌套随机模拟法,该计算方法所需的计算时间降低了约三分之二,并且误差率可以控制在3%以内。The implementation of SolvencyⅡof PhaseⅡproject has put forward higher requirements for insurers’risk management,and the calculation of economic capital has received attention from insurers.Nested stochastic simulation method is the traditional method for calculating economic capital,but most companies cannot afford the time cost and computing power required by this method.A neural network algorithm-based economic capital calculation method was proposed.The results show that the computation time is reduced by about two-thirds compared with the nested stochastic simulation method,and the error rate can be controlled within 3%.

关 键 词：神经网络 经济资本 保险 变额年金 

分 类 号：F840.32[经济管理—保险]