Here, we propose a new kernel approach based on the earth mover's distance (EMD) for electroencephalography (EEG) signal classification. The EEG time series are first transformed into histograms in this approach. The distance between these histograms is then computed using the EMD in a pair-wise manner. We bring the distances into a kernel form called kernel EMD. The support vector classifier can then be used for the classification of EEG signals. The experimental results on the real EEG data show that the new kernel method is very effective, and can classify the data with higher accuracy than traditional methods.
ChaovalitwongseWAPardalosPM. On the time series support vector machine using dynamic time warping kernel for brain activity classification. Cybern Syst Anal. 2008;44(1):125–138.
2.
SubasiA. EEG signal classification using wavelet feature extraction and a mixture of expert model. Expert Syst Appl. 2007;32(4):1084–1093.
3.
PolatKGunesS. Artificial immune recognition system with fuzzy resource allocation mechanism classifier, principal component analysis and FFT method based new hybrid automated identification system for classification of EEG signals. Expert Syst Appl. 2008;34(3):2039–2048.
4.
PolatKGunesS. Classification of epileptiform EEG using a hybrid system based on decision tree classifier and fast fourier transform. Appl Math Comput. 2007;187(2):1017–1026.
GargGSinghVGroverMNidhiGupta JRP. Optimal kernel learning for EEG based scoring system, Int J Biol Med Res. 2011;2(4):1220–1225.
7.
BarachantABonnetSCongedoMJuttenC. BCI signal classification using a Riemannian-based Kernel, 20th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012), Bruges, Belgium, 2012.
8.
XuWWuJHuangZGuanC.Kernel based hidden markov model with applications to EEG signal classification. The IASTED International Conference on Biomedical Engineering (BioMED); 2005; Austria.
9.
LimaCAMCoelhoALV, Kernel machines for epilepsy diagnosis via EEG signal classification: a comparative study. Artif Intell Med. 2011;53(2):83–95.
AndrzejakRGLehnertzKRiekeCMormannFDavidPElgerCE.Indications of nonlinear deterministic and finite dimensional structures in time series of brain electrical activity: dependence on recording region and brain state. Phys Rev E. 2001;64:061907.
12.
RubnerYTomasiCGuibasLJ. A metric for distributions with applications to image databases. Proceedings of the IEEE International Conference on Computer Vision; 1998.
13.
LingHOkadaK. An efficient earth mover's distance algorithm for robust histogram comparision. IEEE TransPattern Anal Mach Intell (PAMI). 2007;29(5):840–853.
14.
ChapelleOHaffnerPVapnikVN. Support vector machines for histogram-based image classification. IEEE Trans Neural Netw. 1999;10(5):1055–1065.
15.
AlweinELSchapireRESingerY.Reducing multiclass to binary: a unifying approach for margin classifiers. J Machine Learn Res. 2000;1:113–141.
16.
DebnathRTakahideNTakahashiH. A decision based one-against-one method for multi-class support vector machine. Pattern Appl. 2004;7(2):164–175.
17.
RifkinRKlautauA.In defense of one-vs-all classification. J Machine Learn Res. 2004;5:101–141.
18.
CrammerKSingerY.On the learnability and design of output codes for multiclass problems. Proc Annu Conf Comput Learn Theory.2000;35–46.
19.
PlattJCCristianiniNShawe-TaylorJ.Large Margin DAG's for Multiclass Classification, in Advances in Neural Information Processing Systems. Cambridge, MA: MIT Press; 2000:547–553; chapter 12.
20.
HsuC-WLinC-J. A comparison of methods for multiclass support vector machines. IEEE Trans Neural Netw. 2002;13(2):415–425.
21.
ChangC-CLinC-J.LIBSVM: a library for support vector machines. ACM Trans Intell Syst Technol. 2011;2(3):27–1.
22.
KannathalNChooMLRajendra AcharyaUSadasivanPK. Entropies for detection of epilepsy in EEG. Comput Methods Program Biomed. 2005;80(3):187–194.
23.
GulerIUbeyliED. Adaptive neuro-fuzzy inference system for classification of EEG signals using wavelet coefficients. J Neurosci Methods. 2005;148(2):113–121.
24.
UbeyliEDGulerI. Features extracted by eigenvector methods for detecting variability of EEG signals. Pattern Recognit Lett. 2007;28:592–603.
25.
GulerIUbeyliED. Expert systems for time-varying biomedical signals using eigenvector methods. Expert Syst Appl. 2007;32(4):1045–1058.
