季丹丹 閆在在
摘要:近幾年,針對缺失數(shù)據(jù)的處理這方面的應(yīng)用研究大量涌現(xiàn),使得缺失數(shù)據(jù)下的可靠性理論迅速發(fā)展.而在可靠性試驗和壽命試驗中,截尾方案能在試驗所花費的總時間、單元個數(shù)和基于試驗結(jié)果的統(tǒng)計推斷效率之間取得平衡.在這種情況下,一種自適應(yīng)的截尾方案被提出來,并且被許多專家學(xué)者研究應(yīng)用.因此本文討論,基于自適應(yīng)逐次II型截尾樣本,提出了EIG分布的統(tǒng)計推斷理論等問題.對于未知參數(shù),提出了極大似然估計(MLEs).利用MLEs的漸近正態(tài)性得到參數(shù)的近似置信區(qū)間.并運用一組真實數(shù)據(jù)進行模擬討論.
關(guān)鍵詞:EIG分布;截尾數(shù)據(jù);極大似然估計;自適應(yīng)逐次II型截尾
中圖分類號:O212 文獻標(biāo)識碼:A 文章編號:1673-260X(2019)03-0013-05
1 引言
許多情形下,考慮到費用和時間的原因,壽命測試驗通常在所有測試單元都失敗前終止.這種情況下,人們只能得到部分樣本的失效時間,這些數(shù)據(jù)即為截尾數(shù)據(jù).在過去的50年里,一些專家學(xué)者已經(jīng)在研究和討論基于截尾樣本的參數(shù)統(tǒng)計推斷問題.最常見的截尾方案大體分兩種,I型(定時)截尾和II型(定量)截尾.其中I型截尾表示壽命試驗在規(guī)定的時間T內(nèi)終止,II型截尾則表示壽命試驗在第m次失效時終止,其中m是提前設(shè)定的.逐次II型截尾方案是II型截尾方案的推廣形式,表示假設(shè)有n個單元置于壽命試驗中,而只有m個失效單元被觀測到.在觀測到第一個失效單元時,在剩余的未失效單元中隨機移除R1個單元.同樣的,在觀測到第二個失效時間時,R2個單元被隨機移除.壽命試驗將在m個失效單元都被觀測到終止,最后將Rm=n-R1-R2-…-Rm-1個未失效單元全部移除.產(chǎn)生逐次型截尾樣本數(shù)據(jù)的原因很多,如有些航空航天、核反應(yīng)堆等零部件,其試驗消耗成本過高,為節(jié)約時間和費用,通過檢驗后,人們通常會在未失效的產(chǎn)品中取出一部分作為他用.這樣即節(jié)約了成本又知道了產(chǎn)品的特性.再如,對某些產(chǎn)品進行跟蹤調(diào)查時,出于某些原因,使得一些使用者在某個時間后失聯(lián),因而我們對這批產(chǎn)品也就只掌握了部分數(shù)據(jù).對于逐次截尾的廣泛的回顧與討論,讀者們可以參考Aggarwala(1998)[1]、alakrishnan(2008)[2]、Fernandez(2004)[3]、Soliman(2008)[4]和Chansoo K和Keunhee H(2009)[5].
2 自適應(yīng)逐次II型截尾試驗
Ng et al.[7]提出一個自適應(yīng)逐次II型截尾方案,它是I型截尾和II型逐次截尾的混合,既節(jié)約了試驗成本,又增加了統(tǒng)計分析效率.
6 結(jié)語
本文介紹了截尾樣本的由來及種類,并由廣義逐次II型截尾試驗,引入并闡述了自適應(yīng)逐次II型截尾試驗的實施過程.由于截尾數(shù)據(jù)的廣泛應(yīng)用性,本文基于自適應(yīng)逐次II型截尾樣本,討論了EIG分布所含參數(shù)的極大似然估計和近似置信區(qū)間,并運用真實例子模擬討論.
參考文獻:
〔1〕Aggarwala R., Balakrishnan N.. Some properties of progressive censored order statistics from arbitrary and uniform distributions with applications to inference and simulation[J]. Statist. Plann. Inference, 1998,70(1):35-49.
〔2〕Balakrishnan N., Anna Dembinska. Progressively Type-II right censored order statistics from discrete distributions[J]. Journal of Statistical Planning and Inference,2008,138(4):845–856.
〔3〕Fernandez A. J. On estimating exponential parameters with general type-II progressive censoring[J]. Journal of Statistical Planning and Inference, 2004,121(1):135-147.
〔4〕Soliman, Ahmed A. Estimations for pareto model using general progressive censored data and symmetric loss[J]. Communications in statistics-theory and methods, 2008,37(9):1353-1370.
〔5〕Chansoo K., Keunhee H. Estimation of the scale parameter of the Rayleigh distribution under general progressive censoring[J]. Journal of the Korean Statistical Society, 2009,38(3):239-246.
〔6〕季丹丹.一種拓展的逆高斯分布的性質(zhì)及應(yīng)用[D].內(nèi)蒙古:內(nèi)蒙古工業(yè)大學(xué),2017.
〔7〕D. Kundu, A. Joarder, Analysis of Type-II progressively hybrid censored data[J], Comput. Stat. Data Anal. 2006, (50) 2258–2509.
〔8〕H.K.T. Ng, D. Kundu, P.S. Chan, Statistical analysis of exponential lifetimes under an adaptive Type-II progressive censoring scheme[J], Naval Res. Logist.2009, (56) 687–698.
〔9〕Rezapour M., Alamatsaz M. H. On properties of progressively Type-II censored order statistics arising from dependent and non-identical random variables[J]. Statistical Methodology, 2013,10(1):58-71.
〔10〕Mashail M. AL Sobhi, Ahmed A. Soliman. Estimation for the exponentiated Weibull model with adaptive Type-II progressive censored schemes[J]. Applied Mathematical Modelling, 2016,40(2):1180–1192.
〔11〕Nassar M. Estimation of the inverse Weibull parameters under adaptive type-II progressive hybrid censoring scheme[J]. Journal of Computational and Applied Mathematics,2017,315:228–239.
〔12〕魏宗舒.概率論與數(shù)理統(tǒng)計教程[M].北京:高等教育出版社,2008.
〔13〕N.Balakrishnan, Rita Aggarwala, Progressive Censoring Theory,methods and Applications[M]. Statistics for industry and technology, 1956.
〔14〕Rezaei S, Tahmasbi R, Mahmoodi M. Estimation of P[Y < X] for generalized Pareto distribution [J]. J Statist Plan Inference. 2010,140:480-494.
〔15〕Greene W H. Econometric Analysis: Fourth Edition [C]. Upper Saddle River, NJ. 2000.
〔16〕Alan A. Categorical Data Analysis (2nd Ed.) [J]. Journal of the Royal Statistical Society, 2002, 40(4).
〔17〕Valiollahi R, Asgharzadeh A, Raqab MZ.Estimation of P[Y 〔18〕Saracoglua B, Kinacia I, Kundu D. (2012) On estimation of R=P[Y 〔19〕 Childs A, Chandrasekhar B, Balakrishnan N, Kundu D.Exact inference based on type-I and type-II hybrid censored samples from the exponential distribution[J]. Ann Inst Stat Math 2003,55:319-330. 〔20〕Balakrishnan,Cramer,Kamps. Bounds for Means and Variances of Progressive Type II Censored Order Statistics[J]. Statist Probab. Lett.2001,54,301-315. 〔21〕Balakrishnan,N.,Cramer,E.,Progressive censoring from heterogeneous distributions with applications to robustness[J]. Ann.Inst.Statist.Math.2008,60:151-171. 〔22〕Guilbaud. Exact non-parametric confidence intervals for quantiles with progressive type-II censoring[J].Scand.J. Statist. 2001,28:699-713. 〔23〕Guilbaud O., Exact non-parametric confidence, prediction and tolerance intervals with progressive type-II censoring[J]. Scand. J.Statist.2004,31:265–281. 〔24〕U Balasooriya, N Balakrishnan. Reliability sampling plans for lognormal distribution based on progressively censored Samples[J]. IEEE Trans. Reliab. 2000,49:199–203.