Computational Statistics

随着我们进一步进入数据科学时代，所产生的数据类型正变得越来越复杂。加拿大计算统计研究主席Paul McNicholas正在开发新的计算密集型统计方法，以更好地洞察几种日益复杂的数据类型。特别是，第二届主席任期内的工作将侧重于三个目标。首先，将开发一种方法来有效地处理三路数据中的离群值，例如黑白图像或多变量纵向数据。然后，张量变量分布的工作将完成，从而例如彩色图像可以被有效地聚类。最后，上述所有方法都将作为开放源码软件实施并免费提供。作为实施过程的一部分，将调查替代算法的使用，以便最终发布最实用的软件。

As we move further into the data science era, the data types being produced are becoming increasingly complex. Paul McNicholas, Canada Research Chair in Computational Statistics, is developing new, computationally intensive statistical methods to gain better insight into several increasingly complex data types. In particular, work during the second term of the Chair will focus on three objectives. First of all, an approach will be developed to effectively handle outliers in three-way data, e.g., black and white images or multivariable longitudinal data. Then, work on tensor-variate distributions will be completed so that, e.g., coloured images can be effectively clustered. Finally, all of the aforementioned approaches will be implemented and made freely available as open source software. As part of the implementation process, the use of alternative algorithms will be investigated so that the most practically useful software is ultimately released.