International: Workshop on Computational Methods for Robust Statistics; Vancouver, Canada, May 19-22, 2002.

这项美洲项目奖金将资助在太平洋数学科学研究所和英属哥伦比亚大学统计系举办的关于稳健统计方法的国际研讨会，该研讨会将与2002年稳健统计国际会议（ICORS2002）一起举行。美国方面由肯塔基大学的Arnold J. Stromberg博士组织，加拿大方面由英属哥伦比亚大学的Ruben Zamar博士组织，该研讨会将汇集计算稳健统计领域的研究人员和学生。数学科学部也参与了这次研讨会的支持。稳健统计处理的问题是，许多经典统计程序只适用于来自正态分布的总体的数据。另一方面，稳健的统计程序独立于数据的正态性，甚至当数据包含异常值时，也能很好地工作。在过去的25年里，稳健统计理论取得了巨大的进步，但许多非常有用的程序由于难以计算而没有在实践中使用。改进鲁棒程序的计算方法将是有益的，因为目前的计算方法仅适用于小到中等规模的数据集。异常值更有可能出现在较大的数据集中，因此计算方法的改进将允许稳健的技术应用于更大的数据集中。讲习班将作为一个论坛，发展和分享关于改进的计算技术的知识，这些技术将允许在实践中使用更可靠的方法。ICORS2002的网页将专门讨论与讲习班有关的讨论，并将在该网页上张贴论文草稿。电子邮件列表将用于通知与会者的重要发展。

0127522StrombergThis Americas Program award will fund an international workshop on robust statistical methods to be held at the Pacific Institute for the Mathematical Sciences and the Department of Statistics of the University of British Columbia in conjunction with the 2002 International Conference on Robust Statistics (ICORS2002). The workshop, which is being organized on the US side by Dr. Arnold J. Stromberg of the University of Kentucky and on the Canadian side by Dr. Ruben Zamar of the University of British Colombia, will bring together researchers and students in computational robust statistics. The Division of Mathematical Sciences is joining in the support of this workshop. Robust statistics deals with the problem that many classical statistical procedures work well only for data that comes from a population that is normally distributed. Robust statistical procedures, on the other hand, work well independently of the normality of the data or even when the data contains outliers. The last 25 years have seen tremendous advancements in the theory of robust statistics, but many very useful procedures are not used in practice due to the fact that they are difficult to compute. Improved computational methods for robust procedures would be beneficial in that current computational methods are only practical for small to moderate sized data sets. Outliers are much more likely to appear in larger data sets and thus improvement in computational methods will allow robust techniques to be applied to much larger data sets. The workshop will serve as a forum to develop and share knowledge on improved computational techniques that will allow more robust methods to be used in practice. A web page on the ICORS2002 will be devoted to discussions related to the workshop and drafts of papers will be posted there. An electronic mailing list will be used to notify participants of important developments.