ICASSP 2017accepted0 citations

Jeffrey's divergence between moving-average and autoregressive models

Leo Legrand, Éric Grivel

Abstract

This paper deals with model comparison based on the Jeffrey's divergence (JD). More particularly, after providing the JD between the joint distributions of k consecutive values of a white noise and the ones of a real moving-average or autoregressive model, the JD between real 1 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">st</sup> -order MA and real 1 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">st</sup> -order AR models is studied. Except when the 1 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">st</sup> MA parameter is equal to 1, we show that, after a transient period, the JD between both models is incremented by a constant value that depends on the model parameters while k is incremented by 1. The JD is hence characterized by this increment and it is not necessary to consider a lot of samples.

BibTeX
@inproceedings{icassp2017_jeffreysdivergen,
  title = {Jeffrey's divergence between moving-average and autoregressive models},
  author = {Leo Legrand and Éric Grivel},
  booktitle = {ICASSP 2017},
  year = {2017}
}
Jeffrey's divergence between moving-average and autoregressive models · ICASSP 2017