Title :
Statistical and spectral properties of irreducible Markov chains
Author :
Anastassopoulos, Vassilis
Author_Institution :
Electron. Lab., Patras Univ., Greece
Abstract :
The spectral and high order statistical characteristics of irreducible Markov chains are studied. A Markov chain, which corresponds to a single-class state-space, is completely determined by its transition matrix. The paper explains analytically how the type of probability density function (PDF), describing the Markov chain, is determined by its transition matrix. Furthermore, it is shown that the correlation properties (power spectrum) and the high order spectra (HOS) of the Markov chain are also expressed by means of specific terms of the transition matrix.
Keywords :
Markov processes; higher order statistics; matrix algebra; probability; signal processing; spectral analysis; HOS; PDF; correlation properties; higher order statistics; irreducible Markov chains; power spectrum; probability density function; spectral properties; statistical signal processing; transition matrix; transition probability; Digital systems; Eigenvalues and eigenfunctions; Laboratories; Physics; Probability density function; Random processes; Random variables; State-space methods; Stochastic processes; Terminology;
Conference_Titel :
Digital Signal Processing, 2002. DSP 2002. 2002 14th International Conference on
Print_ISBN :
0-7803-7503-3
DOI :
10.1109/ICDSP.2002.1028274