Toward a unified statistical inference by nonlinear-type time-frequency analysis for nonstationary time series
- 2021-05-03 (Mon.), 10:30 AM
- Lecture Hall, B1F, Institute of Statistical Science
- The reception will be held at 10:10 at the Lecture Hall, B1F of the Institute of Statistical Science Building
- Prof. Hau-Tieng Wu
- Department of Mathematics, Department of Statistical Science, Duke University
Abstract
Long-term and high-frequency physiological time series provides rich health information. However, to extract useful biorhythm features from these time series for clinical usage, we encounter several challenges — the time series is usually nonstationary and composed of multiple oscillatory components with complicated statistical features, like time varying amplitude, frequency and non-sinusoidal pattern, and the signal quality is often impaired by complicated noise or artifacts. I will discuss recent progress in dealing with this kind of time series by the nonlinear-type time-frequency analysis along with the recent statistical inference results. I will also demonstrate its application to extracting physiological status from the peripheral venous pressure signal during surgery and open problems.