具有线性趋势回归信度模型的自相关性的score检验  

Score Testing of autocorrelation for Regression Credibility Model with Linear Trend

作　　者：江南[1,2] 林金官[2] 

机构地区：[1]徐州师范大学计算机科学与技术学院,江苏徐州221116 [2]东南大学数学系,江苏南京210096

出　　处：《数理统计与管理》2012年第2期290-295,共6页Journal of Applied Statistics and Management

基　　金：江苏省自然科学基金资助项目(BK2008284)

摘　　要：回归信度模型在保险研究中具有重要的作用。本文讨论了具有线性趋势回归信度模型自相关性的score检验问题。首先推导出模型中自相关存在性检验的score检验统计量,然后利用Monte-Carlo方法模拟了此种检验统计量的功效。最后利用文中所得到的检验方法对旅客意外身体伤害保险数据进行了实例分析。The regression credibility models play very important role in study of insurance. In this paper, the score tests of autocorrelation is discussed for the regression credibility model with linear trend. Firstly, the score test statistic is derived for the existence of autocorrelation. Then the power simulation of the statistics is given by using the Monte-Carlo method, which shows that the test statistics has the excellent power. Finally, the obtained test method is illustrated by the private passenger automobile data.

关 键 词：回归信度模型 自相关性 SCORE检验 功效模拟 

分 类 号：O212.1[理学—概率论与数理统计]