Diagnostic Monitoring of High-dimensional Networked Systems (with Applications in Manufacturing and Healthcare System)

Abstract

Rapid advances in sensor and information technology have resulted in both spatially and temporally data-rich environment, which creates a intensive need for us to develop novel statistical methods and computationally efficient algorithm to extract intelligent knowledge and informative patterns from these complicated system. For example, smart manufacturing, or it is also called advanced manufacturing technology take the benefit of information, technology and human intelligence to bring about a rapid revolution in the development and application of manufacturing. Another example is that, in health care field, such as Alzheimer's disease, the researchers have acquired a large number of biomarkers from various modalities including genotyping, neuroimaging and clinical assessment. These changes and development always produces a complicated high dimensional networked systems. However, the statistical and computational challenges for addressing these complicated systems lay in their complex structures, such as high dimensionality, hierarchy, multi modality, heterogeneity and data uncertainty. On the other hand, a bunch of recent development in statistic, optimization and machine learning, such as graphical model, dimension reduction and feature screening technology provide more insights and angles to address the problems coming from these complex high dimensional networked system. I depict the development of novel statistical model and computationally efficient algorithm to analysis the high dimensional networked system, and show how these proposed models are applied to real world applications, such as manufacturing system and Alzheimer's Disease.