Key result
All four criteria underestimated the true autoregressive order, leading to the recommendation that a fixed AR order of at least p=16 should be used for spectral analysis of short tachogram segments.
Population
Short segments of tachograms used for heart rate variability analysis and a true autoregressive process of…
Authors
Loading...
Caution in AR order selection via standard criteria for short tachogram spectral analysis; leaves open validation of fixed p≥16 in clinical HRV research.
For spectral analysis of short segments of tachograms in HRV analysis, an autoregressive model order of at least 16 is recommended as standard criteria tend to underestimate the true order.
Boardman et al. (2002) studied Heart rate variability analysis. Four criteria for estimation of optimum model order (Akaike's FPE, AIC, Parzen's CAT, Rissanen's MDL) vs. True AR process of known order p=6 was evaluated on Estimation of the correct order of a true AR process. All four criteria underestimated the true autoregressive order, leading to the recommendation that a fixed AR order of at least p=16 should be used for spectral analysis of short tachogram segments.
Synapse has enriched 4 closely related papers on similar clinical questions. Consider them for comparative context: