随机PP Kolmogorov多维拟合优度检验乃其Bootstrap逼近

RANDOM PP TEST FOR MULTIVXRIATE GOODNESS-OF-FIT AND ITS BOOTSTRAP APPROXIMATION

本文提出了一种基于随机选择投影方向的PP型棉球等高分布族的拟合优度检验，其特点是计算上较通常的PP检验统计量简单．得到了其检验统计量在零假设下的极限分布，讨论了其Bootstrap逼近及逼近的相容性．

In this pepar, a PP test for multivariate goodness-of-fit based on random projection is propossed. The calculation of the test statistic is simpler than that we usually use. The asymptotic distribution of the test statistic under null hypothesis is derived. The Bootstrap approximation and its consistence is given.