复杂数据模型下指数Pareto分布的参数估计

Parameter Estimation of Exponential Pareto Distribution Based on Complex Data Model

摘要: 针对服从指数Pareto分布的逐步II型删失竞争风险模型，在潜在失效时间的寿命分布具有不同参数的情况下，使用Newton-Raphson方法获得了未知参数的极大似然估计；在平方误差损失函数下，使用重要性抽样方法获得了未知参数的贝叶斯估计。最后，使用蒙特卡洛模拟对所提方法的性能进行了评估，结果显示两种方法的估计效果都很好。

Abstract: We consider the parameter estimation problem of the type II progressively censored competing risks model obeying the exponential Pareto distribution. When the life distribution of the potential failure time has different parameters, the maximum likelihood estimation of the unknown parameters is obtained by using the Newton-Raphson method, and the Bayesian estimation of the unknown parameters is obtained by using the importance sampling method under the square error loss function. Finally, the performance of the proposed method is evaluated by Monte Carlo simulation. The results show that the estimation results of the two methods are very good.