Saddlepoint Expansions in Finite Population Problems
研究了从有限总体中无放回随机抽样时样本均值的鞍点近似，通过数值例子展示了高精度，并讨论了向更一般统计量的推广。
Saddlepoint approximations are derived for the sample mean in the case where samples are obtained randomly without replacement from a finite population. Asymptotic properties are investigated. The great accuracy is demonstrated by numerical examples. Extensions to more general statistics are discussed.