RESEARCH OF STATISTICAL INFERENCE PROCESS IN THE HIGH-DIMENSIONAL INFORMATION PROCESSING

该项目的目的是研究和开发与人类活动相关的高维信息处理过程中的一些统计方法。在此过程中，我们特别关注了日期自适应或信息环境自适应推理。此外，在统计数据分析的过程中，我们提供了一些在数据调查和修正中没有看到的富有成效的发现。主要成果包括树结构方法、统计转换、统计图形、数据自适应分布、一些混合模型和比例风险模型的变体等。在项目研究中，我们可以获得以下主要成果和一些富有成效的发现。我们系统地提供了高维数据处理特别是数据科学领域的数据自适应推理过程和综合方法。系统地介绍了统计图形的理论、方法、应用和评价，提出了数据自适应图形，并通过案例研究和仿真对其进行了评价。就医学领域的最新发展和广泛应用提供了不完整数据的分析方法。提出了生存CART和纵向MARS (L-MARS)，并将其应用于医学和制药领域文献中引用的一些数据。

The purpose of the project is to research and develop some statistical methodologies in process of the high-dimensional information processing which connects with human activities. Particularly, in the process we have focused on date-adaptive or information environmental-adaptive inference. Further, in the process of statistical date analysis, we provided some productive findings that were unsighted in date investigations and reifications.Main productive results involved tree structured approaches, statistical transformations, statistical graphics, date adaptive distributions, some mixed models, and variants of the proportional hazard model etc. In the project research, we could obtain the following main results and some productive findings.We have provided the process of date-adaptive inference and the comprehensive approaches systematically in the high-dimensional processing, particularly, fields of the date science.In the statistical graphics, theories, methods, applications and evaluations were systematically provided and the data-adaptive graphics were proposed and evaluated though case-studies and simulations.Analytical methods of incomplete data were provided with respect to recent developments and widely applied in medical fields.Survival CART and longitudinal MARS (L-MARS) were proposed, developed, and applied to some data cited in literatures of medical and pharmaceutical fields.