Abstract
In this contribution, periodic and almost periodic stationary processes are respectively studied in the frequency domain. A relation of proximity is clearly established between one of the spectral tools associated with these processes: the associated random measure. This is a way to consider, for such processes, that the filters resulting from Principal Components Analysis in the frequency domain are close.
Information
Published: 1 January 2018
First available in Project Euclid: 26 September 2019
Digital Object Identifier: 10.16929/sbs/2018.100-04-03
Subjects:
Primary:
60B15
,
60G10
,
60G57
,
60H05
,
62M15
,
62P12
Keywords:
principal components analysis
,
Random measures
,
spectral measures
,
stationary random function
,
stochastic integrals
Rights: Copyright © 2018 The Statistics and Probability African Society
