Workshop on
time-series analysis of noisy data

Workshop on
time-series analysis of noisy data

Mountains in the Lake District

13-15 September 2023, Lancaster, UK

How to distinguish oscillations from noise?

 

MODA (Multiscale Oscillatory Dynamics Analysis) is a numerical toolbox developed by the Nonlinear & Biomedical Physics group at Lancaster University and the Nonlinear Dynamics and Synergetic Group at the Faculty of Electrical Engineering, University of Ljubljana, Slovenia under the supervision of Aneta Stefanovska.

 

MODA is designed for analysing real-life time-series that are assumed to be the output of some a priori unknown non-autonomous dynamical system, and deriving important properties about this dynamical system from the time-series. MODA includes methods both for analysing the recordings of a single signal over time, and for analysing a set of recordings of multiple different signals over time. In particular, it has tools for analysing bivariate time-series consisting of the simultaneous recordings of two different signals over time, with a view to examining possible connections between the two signals.

 

Bringing your own data to the workshop:
During the practical sessions we recommend you to bring your own data (if you do not have suitable data then example time-series will be provided).


We recommend bringing the data on a USB stick, as you can then easily save outputs as well.

For the data to be compatible with MODA they:

  • Should be uniformly sampled, and you must know the sampling rate.
  • Should not contain any NaN/Inf values 
  • If in the form of a .mat file, it should be of variable type double
  • Should be in the form of a matrix (multiple signals) or an array (one signal)
  • Can be in the form of csv files, which are also compatible

 

To download MODA, visit the github page and follow the instructions.

 

Example wavelet transform of a respiration signal

POSTER

Posters can be up to A0 in size, in portrait orientation. They will be discussed in the Poster Session on TBC, as well as during lunch and refreshment breaks.    

ABSTRACT

Abstracts will be presented in a book generated for the conference and should be no longer than a single page. Please submit via email at tsand@lancaster.ac.uk  by the 8th of September. You will be notified promptly of the outcome of your submission.