Tackling EEG signal classification with least squares support vector machines: A sensitivity analysis study

Página do item simplificado
dc.contributor.authorLima C.A.M.
dc.contributor.authorCoelho A.L.V.
dc.contributor.authorEisencraft M.
dc.date.accessioned2024-03-13T01:31:17Z
dc.date.available2024-03-13T01:31:17Z
dc.date.issued2010
dc.description.abstractThe electroencephalogram (EEG) signal captures the electrical activity of the brain and is an important source of information for studying neurological disorders. The proper analysis of this biological signal plays an important role in the domain of brain-computer interface, which aims at the construction of communication channels between human brain and computers. In this paper, we investigate the application of least squares support vector machines (LS-SVM) to the task of epilepsy diagnosis through automatic EEG signal classification. More specifically, we present a sensitivity analysis study by means of which the performance levels exhibited by standard and least squares SVM classifiers are contrasted, taking into account the setting of the kernel function and of its parameter value. Results of experiments conducted over different types of features extracted from a benchmark EEG signal dataset evidence that the sensitivity profiles of the kernel machines are qualitatively similar, both showing notable performance in terms of accuracy and generalization. In addition, the performance accomplished by optimally configured LS-SVM models is also quantitatively contrasted with that obtained by related approaches for the same dataset. © 2010 Elsevier Ltd.
dc.description.firstpage705
dc.description.issuenumber8
dc.description.lastpage714
dc.description.volume40
dc.identifier.doi10.1016/j.compbiomed.2010.06.005
dc.identifier.issn0010-4825
dc.identifier.urihttps://dspace.mackenzie.br/handle/10899/37171
dc.relation.ispartofComputers in Biology and Medicine
dc.rightsAcesso Aberto
dc.subject.otherlanguageEEG signal classification
dc.subject.otherlanguageEpilepsy
dc.subject.otherlanguageKernel functions
dc.subject.otherlanguageLeast squares support vector machines
dc.subject.otherlanguageSensitivity analysis
dc.titleTackling EEG signal classification with least squares support vector machines: A sensitivity analysis study
dc.typeArtigo
local.scopus.citations56
local.scopus.eid2-s2.0-77955580523
local.scopus.subjectBiological signals
local.scopus.subjectCommunication channel
local.scopus.subjectData sets
local.scopus.subjectEEG signal classification
local.scopus.subjectEEG signals
local.scopus.subjectElectrical activities
local.scopus.subjectElectroencephalogram signals
local.scopus.subjectHuman brain
local.scopus.subjectKernel function
local.scopus.subjectKernel functions
local.scopus.subjectKernel machine
local.scopus.subjectLeast Square
local.scopus.subjectLeast squares support vector machines
local.scopus.subjectNeurological disorders
local.scopus.subjectParameter values
local.scopus.subjectPerformance level
local.scopus.subjectSensitivity profiles
local.scopus.subjectSVM classifiers
local.scopus.subjectSVM model
local.scopus.updated2024-05-01
local.scopus.urlhttps://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=77955580523&origin=inward
Arquivos
Coleções
Artigos de periódico