摘要
In this article, we derive a new formula for extreme Student t quantiles. We use the fact that the Student t distribution arises as the limit of a variance-mixture of normals. For the normal distribution there is already a tail quantile formula derived by Reiss (1989). We generalize his procedure and transfer it to our scenario. %26lt;br%26gt;Eventually, we compare the quantile estimates of our formula to those from Gafer and Kafadar (1984), who also derived a Student t quantile formula. Using R to generate a benchmark we find that our method is more accurate for very high quantiles.
- 出版日期2012