【摘要】 讨论了当分布函数F(x)经过-log(-log(F(x)))变换后,可以用多项式逼近,使用非参数分布的泛函估计。建立了分布函数的强相合估计,模拟结果表明估计拟合情况要优于经验分布,较为理想。最后,把-log(-log(F(x)))变换应用于研究保险的缺损数据并取得了很好的结果。更多还原

【Abstract】 The paper discusses the use of functional estimation of nonparametric distribution to establish strong consistent estimation of the distribution function, where the distribution function F(x) after - log( - log (F(x))) conversion is assumed to be approximated with polynomial. Performance of simulative results shows that the estimation is satisfactory and better than experiential distribution. At last, we practice - log( - log (F(x))) conversion into studying the decrement lifetime data in insurance and have a good result.更多还原