徐寶 王宇廷 馬藝光
摘 要 在加權(quán)p,q對稱損失函數(shù)下,對實際中廣泛應(yīng)用的兩參數(shù)Pareto分布,當(dāng)刻度參數(shù)已知時,用參數(shù)估計方法,研究了形狀參數(shù)的最小風(fēng)險同變估計的形式和性質(zhì). 得到了最小風(fēng)險同變估計的一般形式,又經(jīng)由該參數(shù)的廣義Bayes估計,得到了最小風(fēng)險同變估計的精確形式,并證明了這一最小風(fēng)險同變估計具有最小最大性,從而它也是該參數(shù)的最小最大估計,由此將Pareto分布形狀參數(shù)的最小風(fēng)險同變估計、廣義Bayes估計以及最小最大估計聯(lián)系起來.
關(guān)鍵詞 Pareto分布;最小風(fēng)險同變估計;Bayes估計;最小最大估計
中圖分類號 O212.8文獻(xiàn)標(biāo)識碼 A文章編號 1000-2537(2017)06-0076-04
Abstract For the Pareto distribution, in the present work, the form and property of the minimum risk equivariant estimator for shape parameters with known local parameters were investigated under weighted p, q symmetric entropy loss by the method of parameter estimation. The general form of the minimum risk equivariant estimator was obtained, and the exact form of the minimum risk equivariant estimator was found using the general Bayes estimator of shape parameter. The minimaxity of this minimum risk equivariant estimator was proved. The relationships among the general Bayes estimator, the minimum risk equivariant estimator, and the minimax estimator have been established.
Key words Pareto distribution; minimum risk equivariant estimator; Bayes estimator; minimax estimator
3 結(jié)束語
Pareto分布參數(shù)的估計是現(xiàn)代統(tǒng)計文獻(xiàn)中比較常見的研究內(nèi)容,本文基于p, q對稱損失函數(shù),經(jīng)由廣義Bayes估計,在刻度參數(shù)給定條件下,較快捷地給出了形狀參數(shù)的最小風(fēng)險同變估計的精確形式,并證明了它還是該參數(shù)的最小最大估計.本文得到的估計形式更簡捷,可根據(jù)需要靈活調(diào)整待定常數(shù)p, q的值以得到預(yù)期的估計形式.
參考文獻(xiàn):
[1] SAMUEL K, 吳喜之. 現(xiàn)代貝葉斯統(tǒng)計學(xué)[M]. 北京:中國統(tǒng)計出版社,2000.
[2] DIXIT D J, NOOGHABI M J. Efficient estimation in the Pareto distribution with the presence of outliers [J]. Stat Math, 2011,8(4):340-355.
[3] WU J W, LEE W C, CHEN S C. Computational comparison for weighted moments estimators and BLUE of the scale parameter of a Pareto distribution with known shape parameter under type II multiply censored sample [J]. Appl Math Comput, 2006,181(2):1462-1470.
[4] 康會光, 師義民. LINEX 損失下Pareto 分布族參數(shù)的經(jīng)驗Bayes估計[J]. 純粹數(shù)學(xué)與應(yīng)用數(shù)學(xué),2001,17(2):169-174.
[5] 謝天華, 葉 鷹. Pareto分布參數(shù)的漸進(jìn)最優(yōu)與可容許的經(jīng)驗Bayes估計[J]. 應(yīng)用數(shù)學(xué),2006,19(增):237-240.
[6] 侯華蕾,師義民,李豪亮. 雙邊定數(shù)截尾下Pareto分布的可靠性[J].數(shù)理統(tǒng)計與管理,2009,28(5):826-830.
[7] 黃 娟. 右刪失數(shù)據(jù)下Pareto分布參數(shù)經(jīng)驗Bayes漸進(jìn)性質(zhì)[J]. 工程數(shù)學(xué)學(xué)報, 2010,27(5):781-788.
[8] 宋立新,王明秋,王曉光. q-對稱熵?fù)p失函數(shù)下Pareto分布參數(shù)估計[J]. 大連理工大學(xué)學(xué)報, 2011,51(4):616-620.
[9] 王 芳,門 慧. 三參數(shù)廣義帕累托分布的似然矩估計[J]. 數(shù)學(xué)年刊, 2013,34(3):299-312.
[10] 王 娟,徐付霞.Pareto 分布中尺度參數(shù)的區(qū)間估計[J]. 哈爾濱商業(yè)大學(xué)學(xué)報(自然科學(xué)版), 2015,31(5):629-633.
[11] YOGESH M T, SOMESH K, CONSTANTINOS P. Estimating the shape parameter of a Pareto distribution under restrictions[J]. Metrika, 2016,79(1):91-111.
[12] 徐 寶,姜玉秋,滕 飛. 一種加權(quán)對稱損失函數(shù)下一類指數(shù)分布模型參數(shù)的估計[J]. 四川師范大學(xué)學(xué)報(自然科學(xué)版),2011,34(4):484-487.
[13] 周光亞,趙振全,趙 文. 數(shù)理統(tǒng)計(Ⅱ)[M]. 長春:吉林大學(xué)出版社,1988.
[14] BERGER J O. Statistical decision theory and Bayesian analysis. 2nd ed[M]. New York:Springer,1985.
[15] 茆詩松,王靜龍,濮曉龍. 高等數(shù)理統(tǒng)計[M].北京:高等教育出版社,1998.