席吉富,蘇彥玉,肖靜怡,羅子怡,程丹,龍兵
基于下記錄值逆Rayleigh模型的估計及預測
席吉富,蘇彥玉,肖靜怡,羅子怡,程丹,龍兵
(荊楚理工學院 數(shù)理學院,湖北 荊門 448000)
當觀測數(shù)據(jù)是下記錄值時,利用經(jīng)典方法討論了逆Rayleigh分布未知參數(shù)的極大似然估計和一致最小方差無偏估計,進而得到了可靠度及失效率的極大似然估計.取未知參數(shù)的先驗分布為Gamma分布,在平方損失函數(shù)下討論了逆Rayleigh模型的未知參數(shù)、概率密度函數(shù)、累積分布函數(shù)及系統(tǒng)可靠度的Bayes估計,并對元件的剩余壽命進行預測.通過蒙特卡洛模擬研究了估計量的性質(zhì),借助數(shù)值實驗對估計量的值進行計算.
逆Rayleigh分布;下記錄值;先驗分布;Bayes估計;預測區(qū)間
逆Rayleigh分布在可靠性研究領域有著重要的應用,許多受試元件的壽命都可以用逆Rayleigh分布來近似描述.關(guān)于該分布的統(tǒng)計性質(zhì)引起了很多學者的關(guān)注,并產(chǎn)生了一些研究成果[1-4].文獻[5]在左刪失樣本下討論了逆Rayleigh分布未知參數(shù)和加速因子的估計.文獻[6]基于完全樣本探討了逆Rayleigh模型的概率密度函數(shù)和累積分布函數(shù)的估計問題,包括極大似然估計、一致最小方差無偏估計、最小二乘估計及分位數(shù)估計等.文獻[7]利用屏蔽數(shù)據(jù)得到了逆Rayleigh分布未知參數(shù)的估計,并通過隨機模擬進行驗證.文獻[8]在平方誤差和LINEX損失函數(shù)下研究了逆Rayleigh分布的Bayes估計,并進行了數(shù)值模擬.文獻[9]在熵損失函數(shù)下得到了逆 Rayleigh分布形狀參數(shù)的Bayes估計和經(jīng)驗 Bayes估計,并討論了其容許性.文獻[10]基于逐步II型截尾樣本,在三種損失函數(shù)下得到了未知參數(shù)的Bayes估計和區(qū)間估計,并給出了數(shù)值模擬.



根據(jù)定義,記錄統(tǒng)計量序列可以被認為是樣本的特殊次序統(tǒng)計,其大小由觀測值和出現(xiàn)順序決定.記錄值在工程、天氣、壽命試驗、體育和經(jīng)濟等方面都有著十分重要的應用.對記錄值的統(tǒng)計研究始于Chandler,現(xiàn)在已經(jīng)向不同的方向發(fā)展.文獻[11]在下記錄值樣本下討論了逆Rayleigh模型未知參數(shù)、可靠度及失效率的估計,并對未來的記錄值進行預測.文獻[12]基于下記錄值探討了未知參數(shù)的極大似然估計和Bayes估計問題,利用Bayes方法得到了未來記錄值的預測.另外,文獻[13-15]也使用不同的方法討論了未來記錄值的預測問題,文獻[16-18]在記錄值樣本下用Bayes方法研究了模型參數(shù)的估計問題.
本文在下記錄值樣本下研究逆Rayleigh模型的概率密度函數(shù)、累積分布函數(shù)及系統(tǒng)可靠度的Bayes估計,并對元件的剩余壽命進行預測.

將式(1)(2)代入式(5)中,可以得到







證明 根據(jù)式(8)(9),可得


證畢.

而并聯(lián)系統(tǒng)的可靠度可以表示為

根據(jù)式(10)(11),可得到定理2.
證明 根據(jù)式(9),則可得


證畢.


表1 估計的均值及均方誤差


表2 下記錄值樣本

表3 點估計和預測區(qū)間
本文根據(jù)下記錄值樣本,分別利用經(jīng)典方法和Bayes方法討論了逆Rayleigh模型中未知參數(shù)、概率密度函數(shù)、累積分布函數(shù)及系統(tǒng)可靠度的估計.文中也對元件剩余壽命進行了預測,借助文獻[11]中的數(shù)值實例對文中的一些估計量進行了計算.另外,記錄值樣本也可以應用于其他的統(tǒng)計推斷問題.
[1] 何輝,周娜,張瑞明.數(shù)據(jù)缺失下逆瑞利分布可靠度的貝葉斯估計[J].統(tǒng)計與決策,2015(10):69-71.
[2] 龍兵,張忠占.Ⅰ型雙刪失下逆Rayleigh分布的統(tǒng)計分析[J].數(shù)學的實踐與認識,2019,49(6):199-207.
[3] Dey S.Bayesian estimation of the parameter and reliability function of an inverse Rayleigh distribution[J].Malaysian Journal of Mathematical Sciences,2012,6(1):113-124.
[4] Srinivasa Rao G,Kantam R R L,Rosaiah K,et al.Estimation of stress strength reliability from inverse Rayleigh distribution[J].Journal of Industrial and Production Engineering,2013,30(4):256-263.
[5] 龍兵,張忠占.左刪失恒定應力部分加速壽命試驗下逆Rayleigh分布的參數(shù)估計[J].浙江大學學報(理學版),2020,47(3):315-321.
[6] Maleki J F,Zare K,Deiri E.Efficient estimation of the PDF and the CDF of the inverse Rayleigh distribution[J].Journal of Statistical Computation and Simulation,2018,88(1):75-88.
[7] Panwar M S,Sudhir B A,Bundel R,et al.Parameter estimation of inverse Rayleigh distribution under competing risk model for masked data[J].Journal of Institute of Science and Technology,2015,20(2):122-127.
[8] 李蘭平.平方誤差和LINEX損失函數(shù)下逆Rayleigh分布參數(shù)的經(jīng)驗Bayes估計[J].統(tǒng)計與決策,2013(1):81-83.
[9] 王琪,陽連武.熵損失函數(shù)下逆Rayleigh分布形狀參數(shù)的估計[J].赤峰學院學報(自然科學版),2011,27(7):13-14.
[10] 李中恢.逐步Ⅱ型壽命下逆Rayleigh分布參數(shù)的估計[J].宜春學院學報,2012,34(8):35-37.
[11] Shawky A I,Badr M M.Estimations and prediction from the inverse Rayleigh model based on lower record statistics[J].Life Science Journal,2012,9(2):985-990.
[12] Soliman A,Amin E A,Abd-El Aziz A A.Estimation and prediction from inverse Rayleigh distribution based on lower record values[J].Applied Mathematical Sciences,2010,4(62):3057-3066.
[13] Sana S C,F(xiàn)aizan M.Bayesian estimation using Lindley′s approximation and prediction of generalized exponential distribution based on lower record values[J].Journal of Statistics Applications and Probability,2021,10(1):61-75.
[14] Grigoriy V,Udo K.Maximum observed likelihood prediction of future record values[J].Test,2020,29(4):1-26.
[15] LEE J,SONG J J,KIM Y,et al.Estimation and prediction of record values using Pivotal quantities and copulas[J].Mathematics, 2020,8(10):1678-1682.
[16] Kzlaslan F.E-Bayesian estimation for the proportional hazard rate model based on record values[J].Communications in Statistics: Simulation & Computation,2019,48(2):350-371.
[17] WANG Liang,SHI Yimin,YAN Weian.Inference for Gompertz distribution under records[J].Journal of Systems Engineering and Electronics,2016,27(1):271-278.
[18] 龍沁怡,徐麗平.基于上記錄值Lomax分布的統(tǒng)計推斷[J].江西師范大學學報(自然科學版),2023,47(2):216-220.
[19] 陽連武,黃偉凡.基于記錄值樣本的Topp-Leone分布參數(shù)的Bayes估計[J].宜春學院學報,2017,39(9):5-7.
Estimation and prediction of the inverse Rayleigh model based on lower record values
XI Jifu,SU Yanyu,XIAO Jingyi,LUO Ziyi,CHENG Dan,LONG Bing
(School of Mathematics and Physics,Jingchu University of Technology,Jingmen 448000,China)
When the observed data are the lower record values,the maximum likelihood estimation and uniform minimum variance unbiased estimation of the unknown parameter for the inverse Rayleigh distribution are discussed by using the classical method,and the maximum likelihood estimates of the reliability and failure rate are obtained.Taking Gamma distribution as prior distribution of the unknown parameter,the Bayesian estimates of the unknown parameter,probability density function,cumulative distribution function and system reliability of the inverse Rayleigh model are discussed under the squared loss function,and the remaining lifetime of the component is predicted.The properties of the estimators are studied through Monte-Carlo simulation.Finally,numerical examples are used to calculate the values of the estimators.
inverse Rayleigh distribution;lower record values;prior distribution;Bayesian estimation;prediction interval
O213.2
A
10.3969/j.issn.1007-9831.2024.01.004
1007-9831(2024)01-0012-06
2023-05-17
荊楚理工學院教育教學研究項目(JX2022-011);2023年湖北省大學生創(chuàng)新創(chuàng)業(yè)訓練計劃項目(S202311336055)
席吉富(2003-),男,湖北恩施人,在讀本科生.E-mail:256829334@qq.com
龍兵(1973-),男,湖北荊門人,教授,碩士,從事概率統(tǒng)計研究.E-mail:qh-longbing@163.com