Beschreibung
The problem of modeling multiple time series in the spectral domain arisesnaturally in fields where information about frequency behavior is relevant
and several signals are recorded concurrently. For example, multichannel
electroencephalography (EEG) records measurements of electrical potential
fluctuations at multiple locations on the scalp of a subject. I will present a
hierarchical Bayesian modeling approach to spectral density estimation for
multiple time series, where the log-periodogram of each series is modeled as
a mixture of Gaussian distributions with frequency-dependent weights and
mean functions. The implied model for each log-spectral density is a mixture
of mean functions with frequency-dependent weights. In addition to accommodating
flexible spectral density shapes, a practically important feature
of the proposed formulation is that it allows for ready posterior simulation
through a Gibbs sampler with closed form full conditional distributions for
all model parameters. I will show results for multichannel electroencephalographic
recordings, which provide the key motivating application for the proposed
methodology. I will then present some extensions for non-stationary
time series.
Zeitraum | 13 Juni 2017 → 15 Juni 2017 |
---|---|
Ereignistitel | BISP10 |
Veranstaltungstyp | Keine Angaben |
Bekanntheitsgrad | International |