Summary
International Symposium on Nonlinear Theory and its Applications
2005
Session Number:3-1-4
Session:
Number:3-1-4-1
A new approach to mutual information between pairs of time series
Andreas Galka, Tohru Ozaki, Okito Yamashita,
pp.626-629
Publication Date:2005/10/18
Online ISSN:2188-5079
DOI:10.34385/proc.40.3-1-4-1
PDF download (301.7KB)
Summary:
We study the estimation of mutual information between pairs of simultaneous time series, taking into account temporal correlations within the time series. It is shown that an intimate relationship exists between parametric model fitting by MaximumLikelihood and estimation of mutual information. As a result it becomes possible to detect weak correlations within short spatiotemporal data sets, such as provided by the fMRI technique in neuroscience.