PMID- 20621291
OWN - NLM
STAT- MEDLINE
DCOM- 20101129
LR  - 20100816
IS  - 1879-0534 (Electronic)
IS  - 0010-4825 (Linking)
VI  - 40
IP  - 8
DP  - 2010 Aug
TI  - Tackling EEG signal classification with least squares support vector machines: a 
      sensitivity analysis study.
PG  - 705-14
LID - 10.1016/j.compbiomed.2010.06.005 [doi]
AB  - The 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.
CI  - Copyright 2010 Elsevier Ltd. All rights reserved.
FAU - Lima, Clodoaldo A M
AU  - Lima CA
AD  - Graduate Program in Electrical Engineering, School of Engineering, Mackenzie 
      Presbyterian University, Rua da Consolacao, 896, 01302-907 Sao Paulo, SP, Brazil. 
      moraes@mackenzie.br
FAU - Coelho, Andre L V
AU  - Coelho AL
FAU - Eisencraft, Marcio
AU  - Eisencraft M
LA  - eng
PT  - Journal Article
PT  - Research Support, Non-U.S. Gov't
PL  - United States
TA  - Comput Biol Med
JT  - Computers in biology and medicine
JID - 1250250
SB  - IM
MH  - *Algorithms
MH  - *Artificial Intelligence
MH  - Brain/physiology/physiopathology
MH  - Electroencephalography/*methods
MH  - Epilepsy/physiopathology
MH  - Humans
MH  - Least-Squares Analysis
MH  - *Signal Processing, Computer-Assisted
EDAT- 2010/07/14 06:00
MHDA- 2010/12/14 06:00
CRDT- 2010/07/13 06:00
PHST- 2009/09/01 00:00 [received]
PHST- 2010/04/19 00:00 [revised]
PHST- 2010/06/11 00:00 [accepted]
PHST- 2010/07/13 06:00 [entrez]
PHST- 2010/07/14 06:00 [pubmed]
PHST- 2010/12/14 06:00 [medline]
AID - S0010-4825(10)00087-9 [pii]
AID - 10.1016/j.compbiomed.2010.06.005 [doi]
PST - ppublish
SO  - Comput Biol Med. 2010 Aug;40(8):705-14. doi: 10.1016/j.compbiomed.2010.06.005.