Asymptotic Kullback-Leibler Increment to Characterize Experiment-induced Stress
Estelle Blons, Éric Grivel, Véronique Deschodt-Arsac, Véronique Lespinet-Najib
Abstract
In this paper, we first propose to analyze the properties of the Kullback-Leibler divergence between wide-sense stationary random processes that can be modeled by ARMA or ARFIMA processes. This study makes it possible to introduce a new feature useful to compare two random processes and called "the asymptotic KL increment". The latter depends on various parameters such as the inverse filters associated to the random processes. An interpretation of "the asymptotic KL increment" is also given. Then, we propose to use it in order to compare the inter-beat intervals, which characterize the cardiac rhythm, when the subjects are either in a calm and soothing situation or under stress.
BibTeX
@inproceedings{icassp2019_asymptotickullba,
title = {Asymptotic Kullback-Leibler Increment to Characterize Experiment-induced Stress},
author = {Estelle Blons and Éric Grivel and Véronique Deschodt-Arsac and Véronique Lespinet-Najib},
booktitle = {ICASSP 2019},
year = {2019}
}