Precise Asymptotics in the Law of Logarithm for U-statistics
U-统计量对数律的精确渐近性

Let {Xn; n ≥1} be a sequence of i.i.d. random variables, and Un be a U-statistic based on the symmetric kernel function h(x, y). Set Un =2/n(n-1) ∑1≤i