BEGIN:VCALENDAR VERSION:2.0 X-WR-CALNAME:EventsCalendar PRODID:-//hacksw/handcal//NONSGML v1.0//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/New_York LAST-MODIFIED:20240422T053451Z TZURL:https://www.tzurl.org/zoneinfo-outlook/America/New_York X-LIC-LOCATION:America/New_York BEGIN:DAYLIGHT TZNAME:EDT TZOFFSETFROM:-0500 TZOFFSETTO:-0400 DTSTART:19700308T020000 RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU END:DAYLIGHT BEGIN:STANDARD TZNAME:EST TZOFFSETFROM:-0400 TZOFFSETTO:-0500 DTSTART:19701101T020000 RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU END:STANDARD END:VTIMEZONE BEGIN:VEVENT CATEGORIES:College of Engineering,Thesis/Dissertations DESCRIPTION:Abstract: Modern systems support critical services in transport ation, energy, communication, finance, defense, healthcare, and other doma ins where failures or prolonged disruptions can lead to economic loss, deg raded service, safety risks, and loss of public trust. As these systems be come more complex, interconnected, and data-driven, ensuring that they rem ain reliable during normal operation and resilient under disruptive condit ions become increasingly challenging. Previous research on system reliabil ity has developed statistical and machine learning models for defect predi ction and defect-prone module classification. Similarly, previous research on system resilience has developed metrics and models to assess how syste ms absorb, recover from, and adapt after disruptions. These approaches pro vide important foundations for system assessment, but many remain limited by insufficient treatment of causal and conditional dependencies among sys tem metrics or insufficient flexibility for nonlinear, dynamic, and time-v arying conditions. As a result, existing models may not fully support pred iction and decision-making when defects, covariates, disruptions, and reco very processes evolve over time. To address these limitations, this disser tation develops statistical and machine learning models for system reliabi lity and resilience prediction. For system reliability, this dissertation develops a causal inference-inspired feature selection approach for defect -prone module classification, allowing cause-effect relationships and cond itional dependencies among system metrics to guide the selection of inform ative predictors. For system resilience, this dissertation develops predic tive models for forecasting system performance under dynamic conditions, i ncluding models that account for delayed effects, nonlinear relationships, and time-varying covariates. The proposed models are validated using benc hmark and real-world datasets related to defect-prone module classificatio n and resilience prediction. Goodness-of-fit and classification performanc e measures are used to compare the proposed approaches with established st atistical and machine learning baselines. The results show that incorporat ing causal and conditional dependencies, delayed effects, nonlinear relati onships, and covariate behavior improves predictive performance across the problems considered. These findings demonstrate that the proposed models can support practical decisions such as identifying high-risk modules, pla nning recovery actions, and forecasting long-term system performance. Over all, this dissertation provides flexible and interpretable modeling approa ches that advance system reliability and  resilience assessment for compl ex systems operating in uncertain and changing environments. ADVISOR(S):  Dr. Lance Fiondella, Department of Electrical & Computer Engineering (lf iondella@umassd.edu) COMMITTEE MEMBERS:       Dr. Alfa Heryudono, Depa rtment of Mathematics Dr. Gokhan Kul, Department of Computer & Information Science Dr. Ruolin Zhou, Department of Electrical & Computer Engineering Dr. Hong Liu, Department of Electrical & Computer Engineering                                            NOTE: All EAS Studen ts are ENCOURAGED to attend.\nEvent page: https://www.umassd.edu/events/cm s/9-25-26-eas-doctoral-dissertation-defense-by-fatemeh-salboukh.php X-ALT-DESC;FMTTYPE=text/html:

Abstract:

\n

Modern system s support critical services in transportation\, energy\, communication\, f inance\, defense\, healthcare\, and other domains where failures or prolon ged disruptions can lead to economic loss\, degraded service\, safety risk s\, and loss of public trust. As these systems become more complex\, inter connected\, and data-driven\, ensuring that they remain reliable during no rmal operation and resilient under disruptive conditions become increasing ly challenging. Previous research on system reliability has developed stat istical and machine learning models for defect prediction and defect-prone module classification. Similarly\, previous research on system resilience has developed metrics and models to assess how systems absorb\, recover f rom\, and adapt after disruptions. These approaches provide important foun dations for system assessment\, but many remain limited by insufficient tr eatment of causal and conditional dependencies among system metrics or ins ufficient flexibility for nonlinear\, dynamic\, and time-varying condition s. As a result\, existing models may not fully support prediction and deci sion-making when defects\, covariates\, disruptions\, and recovery process es evolve over time. To address these limitations\, this dissertation deve lops statistical and machine learning models for system reliability and re silience prediction. For system reliability\, this dissertation develops a causal inference-inspired feature selection approach for defect-prone mod ule classification\, allowing cause-effect relationships and conditional d ependencies among system metrics to guide the selection of informative pre dictors. For system resilience\, this dissertation develops predictive mod els for forecasting system performance under dynamic conditions\, includin g models that account for delayed effects\, nonlinear relationships\, and time-varying covariates.

\n

The proposed models are validated using b enchmark and real-world datasets related to defect-prone module classifica tion and resilience prediction. Goodness-of-fit and classification perform ance measures are used to compare the proposed approaches with established statistical and machine learning baselines. The results show that incorpo rating causal and conditional dependencies\, delayed effects\, nonlinear r elationships\, and covariate behavior improves predictive performance acro ss the problems considered. These findings demonstrate that the proposed m odels can support practical decisions such as identifying high-risk module s\, planning recovery actions\, and forecasting long-term system performan ce. Overall\, this dissertation provides flexible and interpretable modeli ng approaches that advance system reliability and  resilience assessment for complex systems operating in uncertain and changing environments.

\ n

ADVISOR(S):  Dr. Lance Fiondella\, Department of Electrical & Compute r Engineering (lfiondella@umassd.edu)

\n

COMMITTEE MEMBERS:      < /p>\n

\n

NOTE: All EAS Students are ENCOURAGED to attend.

Event page: https://www.umassd.edu/events/cms/9-25-26-eas-doct oral-dissertation-defense-by-fatemeh-salboukh.php

DTSTAMP:20260909T180005 DTSTART;TZID=America/New_York:20260925T120000 DTEND;TZID=America/New_York:20260925T140000 LOCATION:LIB 426 SUMMARY;LANGUAGE=en-us:EAS Doctoral Dissertation Defense by Fatemeh Salbouk h UID:bbf85ea9136521b9683e94931a839bc3@www.umassd.edu END:VEVENT END:VCALENDAR