Skip to content

Self-Tuning Optimized Kalman Filtering (STOK) + DyNet simulation + connectivity metrics

License

Notifications You must be signed in to change notification settings

PscDavid/dynet_toolbox

Repository files navigation

dynet_toolbox

The complete collection of functions and scripts for estimating and simulating time-varying Multivariate Autoregressive processes (tv-MVAR) by means of Kalman filtering and Self-Tuning Optimized Kalman filtering.

The toolbox includes:

  • One demo (please refer to the file dynet_demo01.m for a brief tutorial)
  • Four folders:
    • 'statespace'

      • dynet_SSM_KF.m implements the Kalman filter for state-space modeling of physiological time series
      • dynet_SSM_STOK.m implements the Self-Tuning Optimized Kalman filter
      • dynet_SSM_siSTOK.m STOK filter with structural priors
    • 'connectivity'

      • dynet_ar2pdc.m estimates the tv PDC from tv-AR coefficients
      • dynet_connplot.m displays connectivity matrices (function of time and frequency) for each combination of signals
      • dynet_parpsd.m estimates the AR coefficients in the frequency domain and the parametric power spectral density of the input signals
    • 'simulation'

      • dynet_sim.m is the simulation framework for tv-MVAR generated surrogate time series
      • review.m displays the 1) structural adjacency matrix, 2) the functional adjacency matrix, 3)surrogate time-series in the time domain,
      1. the power spectral density of surrogate time-series
    • 'utilities' contains all the invoked functions to let all the above listed functions/scripts to properly work

About

Self-Tuning Optimized Kalman Filtering (STOK) + DyNet simulation + connectivity metrics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published