Vigas Valdemiro P, Ortega Edwin M M, Cordeiro Gauss M, Suzuki Adriano K, Silva Giovana O
Department of Exact Sciences, University of São Paulo, Piracicaba, SP, Brazil.
Institute of Mathematics, Federal University of Mato Grosso do Sul, Campo Grande, MS, Brazil.
J Appl Stat. 2021 May 3;49(11):2805-2824. doi: 10.1080/02664763.2021.1922994. eCollection 2022.
The work proposes a new family of survival models called the Odd log-logistic generalized Neyman type A long-term. We consider different activation schemes in which the number of factors has the Neyman type A distribution and the time of occurrence of an event follows the odd log-logistic generalized family. The parameters are estimated by the classical and Bayesian methods. We investigate the mean estimates, biases, and root mean square errors in different activation schemes using Monte Carlo simulations. The residual analysis via the frequentist approach is used to verify the model assumptions. We illustrate the applicability of the proposed model for patients with gastric adenocarcinoma. The choice of the adenocarcinoma data is because the disease is responsible for most cases of stomach tumors. The estimated cured proportion of patients under chemoradiotherapy is higher compared to patients undergoing only surgery. The estimated hazard function for the chemoradiotherapy level tends to decrease when the time increases. More information about the data is addressed in the application section.
这项工作提出了一个新的生存模型族，称为奇数对数逻辑广义奈曼A型长期模型。我们考虑了不同的激活方案，其中因素的数量具有奈曼A型分布，事件发生的时间遵循奇数对数逻辑广义族。参数通过经典方法和贝叶斯方法进行估计。我们使用蒙特卡罗模拟研究了不同激活方案下的均值估计、偏差和均方根误差。通过频率主义方法进行的残差分析用于验证模型假设。我们说明了所提出模型对胃腺癌患者的适用性。选择腺癌数据是因为该疾病是大多数胃肿瘤病例的病因。与仅接受手术的患者相比，接受放化疗的患者的估计治愈比例更高。随着时间的增加，放化疗水平的估计风险函数趋于下降。应用部分将介绍有关数据的更多信息。
