摘要
探讨了独立样本情形下U-统计量的分布的渐近展开,在较一般的条件下证明其Edgeworth展开的余项之误差可达到o（N-1/2）,并构造精度为o（n-1/2）的随机加权逼近.
The asymptotic expansion for the distribution of a U - statistic is considered under assumptions that the samples are independently but not identically distributed, and its Edge-worth expansion is proved to hold with remainder o ( n -1/2 ) under comparatively general conditions while a random weighting approximation to the distribution of interest with accuracy of o (n -1/2) is constructed.
出处
《华南师范大学学报（自然科学版）》
CAS
2002年第4期21-24,共4页
Journal of South China Normal University(Natural Science Edition)
