Collaborative Research Team Projects
Modern Spectrum Methods in Time Series Analysis: Physical Science, Environmental Science and Computer Modeling
This project explores modern spectrum methods in time series analysis, with applications in physical science, environmental science and computer modeling.
Research Category: Information Sciences
Why Study Time Series?
Time is an important factor in the recording of data. In a time series, where a sequence of data points are collected over an interval of time, time is the significant variable.
Natural time series in geophysics and solar physics often have complex stochastic structures and spectra with “many lines”. These are not generally captured by low-order parametric models.
“Many line” problems were first discovered while searching for causes of problems in engineering systems. These were associated with the active solar maximum around 1990, leading to advances in robust spectral estimation.
The large number of periodic components in time series highlights the importance of modeling for physical understanding and for careful statistical characterization.
Areas of Exploration
Time Series Analysis
Includes data analysis of time series using multi-taper and related methods. In addition to extracting information from natural time series, spectral estimation methods will be explored for the output of multi-fidelity computer models.
Includes envisaging applications to models of seismic “noise” background, solar gravity modes, environmental solar effects, pollutants and meteorological phenomena.
Modern Spectrum Methods
Includes developing methods to model processes with features such as: nearly periodic components, nonlinear coupling, non-stationarity and non-Gaussian distributions. This is done to devise appropriate statistical tests for their frequency domain parameters.
Solving Global Challenges
Research Team’s Goal
To develop modern methods for the exploratory data analysis of time series, with the help of international collaborations.
People Behind the Project
David J. Thomson, Team Leader | Queen’s University
Glen Takahara | Queen’s University
Devon Lin | Queen’s University
Keith Thompson | Dalhousie University
Jean-Pierre St-Maurice | University of Saskatchewan
Frank Vernon | University of California, San Diego
Wesley Burr | Health Canada
Alan D. Chave | Woods Hole Oceanographic Institution
Martin Connors | Athabasca University
Colin Farquharson | Memorial University of Newfoundland
Alexander (Sasha) Koustov | University of Saskatchewan
Germán A. Prieto | Massachusetts Institute of Technology
Laureline Sangalli | Royal Military College
Karin Sigloch | Oxford
- David J. Thomson and Frank L. Vernon (11/2016): Some comments on the analysis of ‘’big’’ scientific time series. Proceeding of the IEEE, 104, 22202249.
- Emily Somerset (12/2017): Multitaper methods for cyclostationary feature detection in time series data: application to ACE interplanetary magnetic field data. M.Sc Thesis.
- Claire Botler (12/2017): Seasonality of influenza: an investigation using time series spectral analysis and epidemiological models. M.Sc thesis.
- Jingyi Liang (01/2018): Modeling of dynamic computer experiments with both qualitative and quantitative variables.
- Francois Marshall, Glen Takahara, David Thomson (12/2017): A multitaper test for the detection of non-stationary processes using canonical correlation analysis. Draft for IEEE 2018 signal processing conference.
- Francois Marshall and David J. Thomson: (12/2017): Robust model detection schemes in a non-stationary environment.
- Charlotte Haley and David J. Thomson: (11/2017): On solar modulation of low energy galactic cosmic radiation
Explore More Stories
Find Related Programs
Modern Spectrum Methods in Time Series Analysis: Physical Science, Environmental Science and Computer Modeling is a Collaborative Research Team project. This program tackles complex problems through a three-year research and training agenda.
CANSSI offers approximately $200,000 for this type of project, which requires a team of faculty, postdocs, and students.