BookPDF Available

Estimación Espectral de Señales Biomédicas. Métodos clásicos (FFT) y Paramétricos: Aplicaciones Prácticas con Matlab. Tutorial

Authors:

Abstract

This tutorial, written in Spanish, describes the estimation of spectral components of digital signals using classical non-parametric techniques using the Fourier ttansform (FFT) and parametric approaches based on AutoRegressive (AR) and AutoRegressive Moving-Average (ARMA) models of the signals. The tutorial includes MATLAB routines form Signal Processing and System Identification toolboxes and adapted own MATLAB functions. Several examples of synthetized and real ECG and heart rate variability (HRV) signals ant time series are shown.

Supplementary resource (1)

ResearchGate has not been able to resolve any citations for this publication.
Article
The number of digits it takes to write down an observed sequence x1, …, xN of a time series depends on the model with its parameters that one assumes to have generated the observed data. Accordingly, by finding the model which minimizes the description length one obtains estimates of both the integer-valued structure parameters and the real-valued system parameters.
Article
It is shown that a strongly consistent estimation procedure for the order of an autoregression can be based on the law of the iterated logarithm for the partial autocorrelations. As compared to other strongly consistent procedures this procedure will underestimate the order to a lesser degree.
Article
An analytic determination of the frequency resolution for maximum entropy and conventional Blackman-Tukey spectral estimates is made for the case of known autocorrelation. As the signal-to-noise ratio decreases, the maximum entropy resolution is no better than that achievable by the Blackman-Tukey spectral estimate. The mean resolution of an ensemble of spectra constructed from sampled data sequences agrees with the analytic result.-Author
Article
The quality of estimated AR models depends on the estimation algorithm and on the order selection criterion in the finite sample practice. Samples are considered to be finite if the maximum candidate order for the model is greater than N/10, where N denotes the sample size. The Finite Sample Theory describes the statistics of residuals and predictions for several autoregressive estimation algorithms. This leads to finite sample criteria FSIC and CIC for order selection. Those special criteria are necessary because of the increase of the variance of order selection criteria for higher model orders. Previous classes of asymptotical and finite sample criteria only take into account the expectation of the logarithm of the residual energy. However, the increasing variance of that logarithm causes an undesirable tendency to select very high model orders without the special precautions of CIC or FSIC.