LPC distance measures and statistical tests with particular reference to the likelihood ratio

Souza, Peter de; Thomson, Peter

doi:10.1109/tassp.1982.1163867

Cited by 43 publications

(13 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Both these statistics start to grow when a jump has occured, and again the task of the stopping rule is to decide whether the growth is significant. Some other proposed distance measures, in the context of speech processing, are listed in de Souza and Thomson (1982). These two statistics are evaluated on a number of real speech data sets in Andre-Obrecht (1988) for the growing window approach.…”

Section: The Divergence Testmentioning

confidence: 99%

Adaptive Filtering and Change Detection

Gustafsson¹

2001

602

699

View full text Add to dashboard Cite

Section: The Divergence Testmentioning

confidence: 99%

Adaptive Filtering and Change Detection

Gustafsson¹

2001

602

699

View full text Add to dashboard Cite

“…Several tests have been proposed for use in segmentation algorithms. de Souza and Thomson [1982] derived a test based on a likelihood ratio statistic, which is insensitive to changes in spectral magnitude but sensitive to changes in spectral shape. This test statistic has an asymptotic chi‐square distribution with n degrees of freedom under the null hypothesis of no changes in the AR parameters.…”

Section: Introductionmentioning

confidence: 99%

“…In the first stage, a preliminary boundary is detected after the change point; the optimal boundary is then searched in the second stage. Lovell and Boashash [1987] adopted the segmentation procedure of Appel and Brandt [1983], but replaced the test derived by de Souza and Thomson [1982]. Lovell and Boashash [1987] maintained that variations in the spectral magnitude are largely due to the measuring instrument, so they designed the segmentation algorithm to determine segments according to changes in the spectral shape alone.…”

Section: Introductionmentioning

confidence: 99%

“…Lovell and Boashash [1987] maintained that variations in the spectral magnitude are largely due to the measuring instrument, so they designed the segmentation algorithm to determine segments according to changes in the spectral shape alone. Because the test derived by de Souza and Thomson [1982] is insensitive to changes in spectral magnitude, it is questionable whether the resulting segments from Lovell and Boashash 's [1987] method are still stationary.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Segmentation of temperature microstructure

Chen¹,

Hondzo

Rao

2002

J. Geophys. Res.

View full text Add to dashboard Cite

[1] The so-called indirect method of estimating the kinetic energy dissipation rate, a key parameter in environmental fluid mechanics, involves fitting the observed spectrum to the theoretical Batchelor spectrum. This requires the statistically nonstationary temperature gradient profiles to be split onto statistically stationary segments. This comparative study of segmentation algorithms uses synthetic temperature gradient series and the temperature gradient profiles that are measured in an inland lake. The results of segmentation based on autoregressive (AR) models and wavelet analysis are compared. The bias in estimating the Batchelor spectrum from nonstationary segments, which have changes in the spectral shape or the spectral magnitude, is demonstrated. The estimated spectra of the resulting stationary segments should be stationary in both the shape and the magnitude of the spectra. A modified AR-based test and a proposed wavelet-based test are sensitive to changes in both the spectral shape and the spectral magnitude. An empirical segmentation technique sensitive to changes in the spectral shape only is not recommended for use.

show abstract

“…Simulations in [ 2 ] show that Dds is a powerful, symmetric distance measure and is insensitive to changes in signal energy levels. Dds may be used to replace Dab in the Appel and Brandt algorithm but it does not identify the segment boundaries as precisely [ S I .…”

Section: S(tf)mentioning

confidence: 99%

Segmentation of non-stationary signals with applications

Lovell

Boashash

ICASSP-88., International Conference on Acoustics, Speech, and Signal Processing

View full text Add to dashboard Cite

Non-stationary signals are partitioned into near stationary segments using a modified Appel and Brandt algorithm. The modification requires two spectral distance measures to be used to produce an algorithm which is insensitive to changes in signal energy level which are irrelevant in this application. Performance on real and simulated data is presented. Segmentation has been used to provide an estimator of the Evolutive Spectrum and an application to a noisy communication signal is presented.

show abstract

LPC distance measures and statistical tests with particular reference to the likelihood ratio

Cited by 43 publications

References 14 publications

Adaptive Filtering and Change Detection

Adaptive Filtering and Change Detection

Segmentation of temperature microstructure

Segmentation of non-stationary signals with applications

Contact Info

Product

Resources

About