ICA based on asymmetry

Spurek, Przemysław; Tabor, Jacek; Rola, Przemysław; Ociepka, Michał

doi:10.1016/j.patcog.2017.02.019

Cited by 14 publications

(15 citation statements)

References 68 publications

(89 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is known that ICA can produce not only a valid statistical model of the data, but also sources with physical or physiological meaning. Examples of this are the extraction of physiologically significant patterns for EEG data [24], the extraction of atrial rhythms during heart fibrillation [25], the removal of physiological artifacts from the EEG signal [26], and the similarities between ICA and image processing in the visual cortex [27]. This capability is inherited by the proposed method.…”

Section: Generalized Sequential Icamm (G-sicamm)mentioning

confidence: 99%

“…Typically, each column of the mixing matrix is considered as a spatial pattern ("scalp map") of activation of the sources, with the corresponding source being interpreted as the level of activation of the map during the experiment. Popular applications of ICA on EEG have been to locate dipolar sources in the brain and to remove artifact (non-EEG) sources (see [24,26] and the references within). HMM has also been applied to analyze the event-related dynamics of brain oscillations and the causality of physiological phenomena [37].…”

Section: Sensitivity Analysismentioning

confidence: 99%

See 1 more Smart Citation

Multichannel dynamic modeling of non-Gaussian mixtures

Safont

Salazar

Vergara

et al. 2019

Pattern Recognition

View full text Add to dashboard Cite

This paper presents a novel method that combines coupled hidden Markov models (HMM) and non-Gaussian mixture models based on independent component analyzer mixture models (ICAMM). The proposed method models the joint behavior of a number of synchronized sequential independent component analyzer mixture models (SICAMM), thus we have named it generalized SICAMM (G-SICAMM). The generalization allows for flexible estimation of complex data densities, subspace classification, blind source separation, and accurate modeling of both local and global dynamic interactions. In this work, the structured result obtained by G-SICAMM was used in two ways: classification and interpretation. Classification performance was tested on an extensive number of simulations and a set of real electroencephalograms (EEG) from epileptic patients performing neuropsychological tests. G-SICAMM outperformed the following competitive methods: Gaussian mixture models, HMM, Coupled HMM, ICAMM, SICAMM, and a long short-term memory (LSTM) recurrent neural network. As for interpretation, the structured result returned by G-SICAMM on EEGs was mapped back onto the scalp, providing a set of brain activations. These activations were consistent with the physiological areas activated during the tests, thus proving the ability of the method to deal with different kind of data densities and changing nonstationary and non-linear brain dynamics.

show abstract

Section: Generalized Sequential Icamm (G-sicamm)mentioning

confidence: 99%

Section: Sensitivity Analysismentioning

confidence: 99%

Multichannel dynamic modeling of non-Gaussian mixtures

Safont

Salazar

Vergara

et al. 2019

Pattern Recognition

View full text Add to dashboard Cite

show abstract

“…The algorithm is designed especially for problems with asymmetric sources. Split Gaussian ICA (SgICA) (Spurek et al, 2017) is based on the maximum likelihood estimation. In such a case we search for the coordinate system optimally fitted to data as well as the marginal densities such that the data density factors in the base are the product of marginal densities.…”

Section: Introductionmentioning

confidence: 99%

Independent Component Analysis based on multiple data-weighting

Bedychaj¹,

Spurek²,

Łukasz³

et al. 2019

Preprint

View full text Add to dashboard Cite

Independent Component Analysis (ICA) -one of the basic tools in data analysis -aims to find a coordinate system in which the components of the data are independent. In this paper we present Multiple-weighted Independet Component Analysis (MWeICA) algorithm, a new ICA method which is based on approximate diagonalization of weighted covariance matrices. Our idea is based on theoretical result, which says that linear independence of weighted data (for gaussian weights) guarantees independence. Experiments show that MWeICA achieves better results to most state-of-the-art ICA methods, with similar computational time.

show abstract

“…Another problem with these methods, is that they usually assume that the underlying density is symmetric, which is rarely the case. For weak-kurtosic but skewed sources, such methods could fail [24,25].…”

Section: Introductionmentioning

confidence: 99%

“…In most of the cases Split Generalized Gaussian distribution fits the data with better precision. The results of ICA SG [25] (described in our previous article), NGPP [29] (which use a combination of third and fourth cumulants) and our method ICA SGG are compared in Fig. 1 for the case of image separation (for more detail comparison we refer to Section 4.).…”

Section: Introductionmentioning

confidence: 99%

ICA based on Split Generalized Gaussian

Spurek

Rola

Tabor

et al. 2019

Schedae Informaticae

View full text Add to dashboard Cite

Independent Component Analysis (ICA) is a method for searching the linear transformation that minimizes the statistical dependence between its components. Most popular ICA methods use kurtosis as a metric of independence (non-Gaussianity) to maximize, such as FastICA and JADE. However, their assumption of fourth-order moment (kurtosis) may not always be satisfied in practice. One of the possible solution is to use third-order moment (skewness) instead of kurtosis, which was applied in ICA_SG and EcoICA. In this paper we present a competitive approach to ICA based on the Split Generalized Gaussian distribution (SGGD), which is well adapted to heavy-tailed as well as asymmetric data. Consequently, we obtain a method which works better than the classical approaches, in both cases: heavy tails and non-symmetric data

show abstract

ICA based on asymmetry

Cited by 14 publications

References 68 publications

Multichannel dynamic modeling of non-Gaussian mixtures

Multichannel dynamic modeling of non-Gaussian mixtures

Independent Component Analysis based on multiple data-weighting

ICA based on Split Generalized Gaussian

Contact Info

Product

Resources

About