Study on the theories and the applications of multivariate simultaneous statistical issues

在本研究中，我们从理论统计和应用的角度对同时统计推断问题进行了研究，并相互交流，发展了分析的新方法。这些新的想法和发展发表在2004年和2005年的联合统计会议和2005年由科学研究补助金(B)主办的三个年度研讨会上：“多元统计建模及相关领域”、“多元同时建模及函数数据分析的应用”和“关于多元同时建模和函数数据分析”。多元比较是我们的主要议题之一，因此我们在两次联合统计会议上组织了专题会议，以促进这一领域的研究发展。我们邀请了多重比较领域的领先者本贾米尼教授，并提出了本贾米尼-霍奇伯格插件方法，该方法目前在世界范围内是众所周知的，定理…更具科学性和应用性的统计学家。对FDR技术及其相关领域进行了深入的探讨。在功能数据分析方面，我们获得了观察老年人行走的海量图像数据。行走是非常重要的运动，因为人类有别于其他动物，而衰老可能是由于行走能力的不足。我们的目标是同时实现手臂和腿的多变量运动。这些手臂和腿部的动作通过功能数据分析的框架被简化和表述。在我们的分析中，衰老主要体现在慢走和快走的差异上，这些发现通过函数数据分析技术得到了统计上的证实。同时推理的另一个有趣的应用主题是人脑图像的分析。我们对调整欧拉特性的高斯随机场进行了大量的仿真研究，得出了最大强度级的p值，这对于从噪声fMRI图像数据中评估人脑的激活区域是非常重要的。这些结果发表在上海举行的MCP 2005国际会议上。较少

In this research we studied the issues of simultaneous statistical inference from the view point of theoretical statistics and applications and communicated with each other to develop the new methodologies for the analysis. These new ideas and developments are published in the 2004 and 2005 Joint Statistical Meeting and the 2005 and three annual symposium sponsored by Grant-in-Aid for Scientific Research (B) : "Multivariate statistical modeling and the related area," "Multivariate simultaneous modeling and applications of functional data analysis" and "On multivariate simultaneous modeling and functional data analysis."Multiple comparison is one of the our major topics, and so we organized the special session in the both Joint Statistical Meetings to promote the research development in this field. We invited the Professor Benjamini, who is the leader in the field of multiple comparison and proposed the Benjamini-Hochberg plug-in method, which is now world-wide well-know to both theoret … More ical and applied statisticians. We deeply discussed the FDR techniques and their related area.As for functional data analysis we obtained the huge image data observing the aged persons' walking. Walking is very important moving for human beings are distinguished from other animals and aging maybe come from deficiency of ability of smooth walking. Our aim is in the simultaneous multivariate movements of arms and legs. These movements of arms and legs are simplified and formulated by the framework of functional data analysis. In our analysis aging greatly appears in the differences of slow walks and rapid walks and these findings are assured statistically by functional data analysis techniques.Another interesting applied topic for simultaneous inference is the analysis for human brain images. We conducted the vast simulation studies on the Gaussian random fields for adjustments for Euler characteristic to derive the p-values of the maximum intensity levels, which are very important for evaluating the activating area of human brains form the noisy fMRI image data. These results were published in the international conference MCP 2005 held in Shanghai. Less