Mathematical Sciences: Statistical Graphics: Diagnostics andInference

使用实时，计算密集的图形方法作为统计问题的诊断程序将被探索。高维回归问题中的残差将用受限大旅行的方法来研究。创新的计算程序，如切片回归，以寻找结构将被检查。这个过程是半参数的，而不是非参数的，当满足一些基本假设时，它有望产生有用的结果。此外，将探索高维问题中不确定性的新颖图形总结。

The use of real time, computationally intensive graphical methods as diagnostical procedures in statistical problems will be explored. Residuals in high dimensional regression problems will be studied by means of restricted grand tours. Innovative computational procedures such as slicing regressions to find structure will be examined. This procedure is semiparametric rather than nonparametric and promises to yield useful results when a few basic assumptions are satisfied. Furthermore novel graphical summaries of uncertainty will be explored in high dimensional problems.