Online Program Home
My Program

Abstract Details

Activity Number: 457 - Novel Statistical Approaches to Time Series of Networks
Type: Topic Contributed
Date/Time: Wednesday, July 31, 2019 : 8:30 AM to 10:20 AM
Sponsor: Section on Nonparametric Statistics
Abstract #304407
Title: Quantile-Frequency Analysis and Functional Principal Components for Discriminant Analysis of Time Series
Author(s): Ta-Hsin Li*
Companies: IBM T. J. Watson Research Center
Keywords: quantile periodogram; spectral analysis; time series; classification; functional data analysis; principal component

Quantile periodogram is a recently proposed tool for spectral analysis of time-series data. Derived from trigonometric quantile regression, the quantile periodogram offers a capability of quantile-frequency analysis (QFA) that characterizes the oscillatory behavior of time series around different quantile levels. This talk introduces a QFA-based functional principal component analysis (FPCA) method that extracts useful features for discriminant analysis of time series. A real-world dataset of nondestructive evaluation (NDE) of mechanical systems is used to demonstrate the advantages of the proposed method over the traditional spectral analysis method based on the ordinary periodogram.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program