Metrics
1 Record Views
Thesis
Performance Analysis of Hotelling's T^2 Control Chart Under Missing Data in IoT Streams
University of West Florida Libraries
Master of Science (MS), University of West Florida
2026
Abstract
This document evaluates the robustness of Hotelling’s T^2 multivariate statistical processcontrol chart (MSPC) in the presence of missing data within Internet of Things and environmental sensor networks. While classical T^2 charts are widely used for anomaly detection, their behavior under incomplete data—a common challenge in modern sensor deployments—has been insufficiently characterized.
Using four environmental datasets, the thesis systematically examines how missing observations, reference window degradation, and interpolation-based reconstruction affect false-alarm control (ARL0) and detection performance (ARL1). Results demonstrate that Hotelling’s T^2 exhibits conditional robustness: performance remains reliable up to approximately 20–25% missingness when the reference covariance structure is stable. Beyond this threshold, interpolation alone cannot correct covariance distortion, leading to unstable run-length distributions and inflated false alarms. Seasonal reference extension and hybrid adaptation strategies improve robustness, while datasets with weak inter-variable correlation or nonstationary dynamics are particularly vulnerable.
These findings provide actionable guidance for practitioners deploying T^2 in incomplete-data environments, clarify the limits of classical MSPC, and identify directions for future research, including multivariate imputation, adaptive control limits, and integration with robust or distribution-free monitoring frameworks.
Details
- Title
- Performance Analysis of Hotelling's T^2 Control Chart Under Missing Data in IoT Streams
- Resource Type
- Thesis
- Contributors
- Thomas Reichherzer (Committee Chair)Achraf Cohen (Committee Member)Brian Jalaian (Committee Member)
- Publisher
- University of West Florida Libraries
- Format
- pdf
- Number of pages
- 89
- Copyright
- Permission granted to the University of West Florida Libraries by the author to digitize and/or display this information for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires the permission of the copyright holder.
- Identifiers
- 99381756997606600
- Academic Unit
- Hal Marcus College of Science and Engineering
- Language
- English
- Awarding Institution
- University of West Florida; Master of Science (MS)
- Theses and Dissertations
- Master of Science (MS), University of West Florida