Nonlinear Profile Data Analysis for System Performance Improvement by

ii I dedicate this dissertation to my parents Shahnaz and Kazem for their endless love, and unlimited supports, endurance and encouragement. iii Acknowledgements I would like to sincerely thank my advisor, Professor Judy Jin for her great guidance, innumerous support and encouragement, and for everything she has taught me throughout the period of my Ph.D. study, without which this dissertation would not have been completed. Because of Professor Jin, my dream of accomplishing my Ph.D. and pursuing my career in academia came true. Special thanks go to other members of the committee providing me with the collaboration opportunities and for their great encouragement and support. My special gratitude also goes to my former advisors in my undergraduate and Mirbahador Arianezhad for teaching me the basics of conducting research and for encouraging me to continue my graduate study. Finally, I would like to express my extreme gratitude to my parents for their endless love and supports during whole my life, and to my siblings and friends for their continuous encouragement.). (trace) (trace) (~ t f Σ Σ z  Figure 3-7. Engine head, cross-section view of valve seat pocket, and gap between valve seat and pocket (left panel), valve seat assembly process (right panel). Abstract The rapid development of distributed sensing and computer technologies has facilitated a wide collection of various nonlinear profiles during system operations, thus resulting in a data-rich environment that provides unprecedented opportunities for improving complex system operations. At the same time, however, it raises new research challenges on data analysis and decision making due to the complex data structures of nonlinear profiles, such as high-dimensional and non-stationary characteristics. In this dissertation, for the purpose of system performance improvement, new methodologies are proposed to effectively model and analyze nonlinear profile data. Specifically, three major research tasks are accomplished. First, the problem of informative sensor and feature selection among massive multi-stream sensing signals is discussed. In this research, a new hierarchical regularization approach called hierarchical non-negative garrote (HNNG) is proposed. At the first level, a group non-negative garrote is developed to select important signals, and at the second level, the individual features within each signal are selected using a modified version of non-negative garrote that can guarantee nice properties for the estimated coefficients. Second, a new methodology has been developed to analyze cyclic nonlinear profile signals for fully characterizing process variations and enhancing the fault diagnosis capability. In the …