题目：Uniform Projection Designs
主讲：孙法省
时间：2017年11月10日 14：30-15：30
地点：明德主楼 1030
摘要：
Uniform designs are widely used in both physical and computer experiments. Designs with low discrepancy have good uniformity in the full-dimensional space but can result in poor projections onto lower-dimensional spaces, which is undesirable when only a few factors are active. We propose a new design criterion, called uniform projection criterion, by focusing on projection uniformity. Uniform projection designs generated under the new criterion not only have small correlations between columns, but also have good space-filling properties in terms of distance and uniformity in all dimensions. We further develop some theoretical results and show that maximin $L_1$-equidistant designs are uniform projection designs. The new criterion is a function of the pairwise $L_1$-distances between the rows, so that the new criterion can be computed at no more cost than a design criterion that ignores projection properties. We also construct of a class of uniform projection designs.
简介：
孙法省，2010于南开大学获统计博士学位。2012年在加拿大西蒙弗雷泽大学做博士后，2016年在加州大学洛杉矶分校访问一年。 现为东北师范大学数学与统计学院副教授，博士生导师。研究方向为计算机试验、试验设计、高维数据分析等。现已在统计学期刊 Annals of Statistics、 Journal of American Statistical Association、Biometrika 等期刊上发表论文近二十篇。