Change-Point Detection and Causal Inference for Time Series with Applications in Healthcare
Author(s)
Wei, Song
Advisor(s)
Editor(s)
Collections
Supplementary to:
Permanent Link
Abstract
Explainable prediction algorithms have become increasingly important in automated surveillance systems within the healthcare context, as they offer actionable insights for clinicians on duty to respond to predicted adverse events. In this thesis, I will present a real study on sepsis prediction, and several novel methods motivated by it. Those methods, developed with the help of recent advancements in statistics and optimization, enjoy strong theoretical guarantees and exhibit promising empirical performance. Importantly, with the numerical demonstration on the real data, I hope the developed methods can be extended to a broader range of real applications.
Sponsor
Date
2024-05-10
Extent
Resource Type
Text
Resource Subtype
Dissertation