Abstract On the basis of an interesting and tractable parametric family of distributions, which figures prominently as an empirical model for asymmetric and heavy-tailed data, we introduce a novel parametric regression model that is quite simple and may be very useful to model many types of real data which occurs frequently in practice. The unknown parameters are estimated using the maximum likelihood estimation method, and Monte Carlo experiments indicate that this traditional approach works properly to estimate the unknown parameters. Diagnostic measures (normalized quantile residuals, and global and local influence methods) are also discussed for the new parametric regression model. Empirical applications are considered, and comparisons with three of the most popular existing regression models are made.

中文翻译：

---

偏斜 sinh-arcsinh t 分布的参数回归框架

摘要 基于一个有趣且易于处理的参数分布族，它作为非对称和重尾数据的经验模型突出显示，我们引入了一种新颖的参数回归模型，该模型非常简单并且可能对建模许多类型的数据非常有用。在实践中经常出现的真实数据。未知参数的估计使用最大似然估计方法，蒙特卡罗实验表明这种传统方法可以正确估计未知参数。还讨论了新参数回归模型的诊断措施（归一化分位数残差以及全局和局部影响方法）。考虑了实证应用，并与三个最流行的现有回归模型进行了比较。