ABSTRACT

The aim of this paper is to compare some parameter estimation methods for the three-parameter kappa distribution, which is used in analysing precipitation, wind speed and streamflow data in hydrology. These methods include: method of L-moments (LM), maximum likelihood (ML), maximum product of spacings (MPS), location and scale parameters free maximum likelihood (LSPF), location parameter free maximum likelihood (LPF) and shrinkage estimation (SH). To compare the performance of these methods, some Monte Carlo simulation studies are applied. To demonstrate the usefulness of the three-parameter kappa distribution and its applications to hydrology, statistical analyses of two precipitation datasets are presented.

摘要

本文的目的是比较三参数 kappa 分布的一些参数估计方法，用于分析水文学中的降水、风速和流量数据。这些方法包括：L矩法（LM）、最大似然法（ML）、最大间距乘积（MPS）、位置和尺度参数自由最大似然法（LSPF）、位置参数自由最大似然法（LPF）和收缩估计（ SH）。为了比较这些方法的性能，应用了一些蒙特卡罗模拟研究。为了证明三参数 kappa 分布的有用性及其在水文学中的应用，本文介绍了两个降水数据集的统计分析。