Newer Haidy A
Department of Mathematics, Faculty of Education, Ain Shams University, Cairo, 11511, Egypt.
Sci Rep. 2025 May 30;15(1):19001. doi: 10.1038/s41598-025-01995-z.
This article presents an innovative sampling strategy, ordered moving extremes lower k-record ranked set sampling, designed to enhance parameter estimation and prediction for the generalized exponential distribution. By incorporating k-record values with random sample sizes, we develop maximum likelihood estimation, classical Bayes estimation, and empirical Bayes estimators, leveraging informative priors under balanced loss functions, including balanced squared error and balanced linear exponential. Additionally, we utilize the pivotal prediction method to construct prediction intervals for future observations under double type-II censoring. Extensive simulation studies demonstrate that our approach significantly improves estimation accuracy by achieving lower mean squared errors and reduced bias compared to conventional methods. The efficacy of the proposed sampling method is further validated through its application to real-world medical datasets, showcasing its practical utility in enhancing statistical inferences for lifetime data analysis. The key findings highlight that ordered moving extremes lower k-record ranked set sampling effectively balances efficiency and accuracy, making it particularly well-suited for reliability studies and survival analysis.
本文提出了一种创新的抽样策略，即有序移动极值下k记录排序集抽样，旨在增强广义指数分布的参数估计和预测。通过将k记录值与随机样本量相结合，我们开发了最大似然估计、经典贝叶斯估计和经验贝叶斯估计量，利用平衡损失函数（包括平衡平方误差和平衡线性指数）下的信息先验。此外，我们利用枢轴预测方法在双重II型删失下为未来观测构建预测区间。广泛的模拟研究表明，与传统方法相比，我们的方法通过实现更低的均方误差和减少偏差，显著提高了估计精度。所提出的抽样方法的有效性通过其在实际医学数据集上的应用得到进一步验证，展示了其在增强寿命数据分析的统计推断方面的实际效用。关键发现突出表明，有序移动极值下k记录排序集抽样有效地平衡了效率和准确性，使其特别适合可靠性研究和生存分析。
