1 |
105/2 |
統計系 蔡宗儒 教授於
期刊論文
發佈
Inference of δ = P(X < Y) for Burr XII distributions with record samples
,
[105-2]
:Inference of δ = P(X < Y) for Burr XII distributions with record samples期刊論文Inference of δ = P(X < Y) for Burr XII distributions with record samplesJyun-You Chiang; Nan Jiang; Tzong-Ru Tsai; Y. L. LioBootstrap-t method;Fisher information;Log-logistic distribution;Maximum likelihood estimation;Percentile bootstrap methodCommunications in Statistics - Simulation and Computation 47(3), p.822-838Let X and Y follow independent Burr type XII distributions, which share a common inner shape parameter. The maximum likelihood estimator of the parameter δ = P(X < Y) is studied based on record samples. The existence and uniqueness of the maximum likelihood estimator of δ based on record samples are established. When the inner shape parameter is known, an exact confidence interval of δ is derived; otherwise, the Fisher information matrix and two bootstrap methods are used to obtain three approximate confidence intervals of δ. The performances of the proposed methods are evaluated via Monte Carlo si
|