具有校正项的K分布形状参数的V-估计器

Oliver于1993年提出的K分布形状参数的V-估计器(VE),虽然具有免于求解非线性方程的计算因而估计效率高的优点,但是,其估计精度却低于许多其他矩估计器的估计精度,且有时会出现奇异值的情况.为扬长避短,在对V-估计器的估计偏差进行推导和分析的基础上,通过一系列Monte-Carlo实验,V-估计器改进成本文提出的具有校正项的V-估计器(VCE).VCE克服了V-估计器的上述缺点.仿真实验表明,VCE的估计精度不但显著优于VE,而且在效率和精度上都优于通常被认为是精度最高的矩估计器中的U-估计器.特别是,实验结果显示,VCE更适合于小样本长度下的情况,这个特点使得它更便于实际应用.

V-estimator with corrective term for K-distribution shape parameter

The V-estimator (VE) for K-distribution shape parameter proposed by Oliver in 1993, bears the characteristics of without solving non-linear equations so it has a high estimating efficiency, but the estimating accuracy of it is lower than that of many other moment estimators, sometimes the VE even results in odd value. In order to make the best use of the advantages and bypass the disadvantages, on the basis of derivation and analysis of the VE bias, by means of a set of Monte-Carlo experiments, the V-estimator with corrective term (VCE) was discussed, which overcomes the shortcomings above of the VE. Simulation results show that not only the estimating accuracy of the VCE is significantly superior to the VE, but also, on both of estimating accuracy and efficiency, to the U-estimator considered as the moment estimator with the highest accuracy usually. Especially, experiment results demonstrate that the VCE is better suited to performing in the case of small samples, this feature makes it possible that the VCE is more applicable to the practice.