基于广义线性回归模型的医疗保险住院费用预测  

Prediction of Hospitalization Expense in Medical Insurance on the Bases of Generalized Linear Model

作　　者：西日古力[1] 吴黎军[1] 

机构地区：[1]新疆大学数学与系统科学学院,乌鲁木齐830001

出　　处：《数理医药学杂志》2011年第3期255-257,共3页Journal of Mathematical Medicine

摘　　要：在医疗保险基金运行中,住院费用支出是一个重要方面。人们以前常常采用古典线性回归模型对住院费用进行分析,但是古典线性回归模型是建立在一些假设之上的。在非寿险精算中,古典线性回归模型的假设很难得到满足。广义线性模型的误差项服从自然指数分布,而普通线性回归模型的误差项服从正态分布,这样就使得广义线性模型更加灵活,适用范围更加广阔。用广义线性回归模型分别对住院费用进行分析和预测。Hospitalization expense is one of the most important parts of the operational status of pooling funds of basic medical insurance.People used to analyze hospitalization expense with general linear models,but general linear model is built on some hypothesis.In non-life insurance,the hypothesis of general linear model is hard to be satisfied.Error terms of generalized linear models are usually allowed exponential family of distributions and error terms of general linear model are usually allowed normal distribution,so generalized linear model is more flexible than general linear model.generalized linear model have wide implicational fields.The paper analyzed and predicted the hospital expense with generalized linear model.

关 键 词：广义线性回归模型 医疗保险 指数分布族 

分 类 号：R195.1[医药卫生—卫生统计学]