基于Bootstrap方法的分位数估计

Quantile Estimation Based on Bootstrap Method

分位数广泛应用于人口调查、服装设计、洪水研究、金融风险分析方面。分位数估计的方法基本分为两大类:一类是参数估计,另一类是非参数估计。文章运用非参数Bootstrap分位数估计研究人体测量指标,并与参数分位数估计进行比较。结果表明,利用Bootstrap抽样方法来估计分位数的结果精度较高且更加稳定。

Quantile is widely used in census, clothing design, flood research and financial risk analysis. Quantile estimation methods are basically divided into two categories: one is parametric estimation, and the other is non-parametric estimation. This paper employs non-parametric Bootstrap quantile estimation to study anthropometric indexes and compares it with parametric quantile estimation. The results show that using Bootstrap sampling method has more accurate and stable result of quantile estimation.