基于推广的回归模型的风险管理方法

Risk Management Based on the Generalized Regression Model

首先,利用Choquet积分将多元线性回归模型进行推广;然后,基于推广的非线性回归模型,将若干交互作用的风险因素抽象成一个数学模型,采用定量分析的方法,在给定风险损失和风险发生概率的历史数据的基础上,计算出各个风险在风险集合中的重要程度及其风险之间交互作用的大小,从而进行有效的风险管理.

Firstly, the linear multiregression model is generalized by using the Choquet integral. Then, a mathematical model is constructed to deal with some risks with intersection based on the generalized regression model. We calculate the importance degree of every risk in the risk set and the size of the intersection of risks given the data of the risk losses and the probability that every risk happened.