Dendrite Ellipsoidal Neuron Trained by Stochastic Gradient Descent for Motor Imagery Classification

Fernando Arce; Omar Mendoza-Montoya; Erik Zamora; Javier M. Antelis; Humberto Sossa; Jessica Cantillo-Negrete; Ruben I. Carino-Escobar; Luis G. Hernández; Luis Eduardo Falcón

doi:10.1007/978-3-030-21077-9_8

Dendrite Ellipsoidal Neuron Trained by Stochastic Gradient Descent for Motor Imagery Classification

Fernando Arce, Omar Mendoza-Montoya, Erik Zamora, Javier M. Antelis, Humberto Sossa, Jessica Cantillo-Negrete, Ruben I. Carino-Escobar, Luis G. Hernández, Luis Eduardo Falcón

Centro de Investigación en Computación (CIC)

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

1 Scopus citations

Abstract

Dendrite ellipsoidal neurons are a novel and different alternative for classification tasks, giving competitive results compared with typical classification methods. Based on k-means++ algorithm, the network allows each dendrite to build a hyperellipsoidal in order to assign each incoming pattern to its respective C class. The main disadvantage of this training algorithm is the lack of accuracy in high dimensional datasets. In this research, we solved this problem by training the dendrite ellipsoidal neuron using stochastic gradient descent. Furthermore, electroencephalography data were acquired during two mental conditions (imaginary movements of the left and right hand) in order to test the new training algorithm. The proposed algorithm outperformed the accuracy acquired by a dendrite ellipsoidal neuron based on k-means++ obtaining 76.02% and 62.77%, respectively. Also, the algorithm was compared with multilayer perceptrons and support vector machines which are some of the most common classifiers used to detect motor-related information in brain signals. These achieved an accuracy of 72.38% and 65.81%, respectively.

Original language	English
Title of host publication	Pattern Recognition - 11th Mexican Conference, MCPR 2019, Proceedings
Editors	Jesús Ariel Carrasco-Ochoa, José Francisco Martínez-Trinidad, José Arturo Olvera-López, Joaquín Salas
Publisher	Springer Verlag
Pages	80-88
Number of pages	9
ISBN (Print)	9783030210762
DOIs	https://doi.org/10.1007/978-3-030-21077-9_8
State	Published - 2019
Event	11th Mexican Conference on Pattern Recognition, MCPR 2019 - Querétaro, Mexico Duration: 26 Jun 2019 → 29 Jun 2019

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11524 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	11th Mexican Conference on Pattern Recognition, MCPR 2019
Country/Territory	Mexico
City	Querétaro
Period	26/06/19 → 29/06/19

Keywords

Dendrite Ellipsoidal Neuron
Electroencephalography
Motor Imagery
Multilayer perceptrons
Stochastic Gradient Descent
Support vector machines
k-means++

Access to Document

10.1007/978-3-030-21077-9_8

Cite this

Arce, F., Mendoza-Montoya, O., Zamora, E., Antelis, J. M., Sossa, H., Cantillo-Negrete, J., Carino-Escobar, R. I., Hernández, L. G., & Falcón, L. E. (2019). Dendrite Ellipsoidal Neuron Trained by Stochastic Gradient Descent for Motor Imagery Classification. In J. A. Carrasco-Ochoa, J. F. Martínez-Trinidad, J. A. Olvera-López, & J. Salas (Eds.), Pattern Recognition - 11th Mexican Conference, MCPR 2019, Proceedings (pp. 80-88). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11524 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-21077-9_8

Arce, Fernando ; Mendoza-Montoya, Omar ; Zamora, Erik et al. / Dendrite Ellipsoidal Neuron Trained by Stochastic Gradient Descent for Motor Imagery Classification. Pattern Recognition - 11th Mexican Conference, MCPR 2019, Proceedings. editor / Jesús Ariel Carrasco-Ochoa ; José Francisco Martínez-Trinidad ; José Arturo Olvera-López ; Joaquín Salas. Springer Verlag, 2019. pp. 80-88 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{35b2302770d34074804d24e81bf03229,

title = "Dendrite Ellipsoidal Neuron Trained by Stochastic Gradient Descent for Motor Imagery Classification",

abstract = "Dendrite ellipsoidal neurons are a novel and different alternative for classification tasks, giving competitive results compared with typical classification methods. Based on k-means++ algorithm, the network allows each dendrite to build a hyperellipsoidal in order to assign each incoming pattern to its respective C class. The main disadvantage of this training algorithm is the lack of accuracy in high dimensional datasets. In this research, we solved this problem by training the dendrite ellipsoidal neuron using stochastic gradient descent. Furthermore, electroencephalography data were acquired during two mental conditions (imaginary movements of the left and right hand) in order to test the new training algorithm. The proposed algorithm outperformed the accuracy acquired by a dendrite ellipsoidal neuron based on k-means++ obtaining 76.02% and 62.77%, respectively. Also, the algorithm was compared with multilayer perceptrons and support vector machines which are some of the most common classifiers used to detect motor-related information in brain signals. These achieved an accuracy of 72.38% and 65.81%, respectively.",

keywords = "Dendrite Ellipsoidal Neuron, Electroencephalography, Motor Imagery, Multilayer perceptrons, Stochastic Gradient Descent, Support vector machines, k-means++",

author = "Fernando Arce and Omar Mendoza-Montoya and Erik Zamora and Antelis, {Javier M.} and Humberto Sossa and Jessica Cantillo-Negrete and Carino-Escobar, {Ruben I.} and Hern{\'a}ndez, {Luis G.} and Falc{\'o}n, {Luis Eduardo}",

note = "Publisher Copyright: {\textcopyright} 2019, Springer Nature Switzerland AG.; 11th Mexican Conference on Pattern Recognition, MCPR 2019 ; Conference date: 26-06-2019 Through 29-06-2019",

year = "2019",

doi = "10.1007/978-3-030-21077-9_8",

language = "Ingl{\'e}s",

isbn = "9783030210762",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "80--88",

editor = "Carrasco-Ochoa, {Jes{\'u}s Ariel} and Mart{\'i}nez-Trinidad, {Jos{\'e} Francisco} and Olvera-L{\'o}pez, {Jos{\'e} Arturo} and Joaqu{\'i}n Salas",

booktitle = "Pattern Recognition - 11th Mexican Conference, MCPR 2019, Proceedings",

address = "Alemania",

}

Arce, F, Mendoza-Montoya, O, Zamora, E, Antelis, JM, Sossa, H, Cantillo-Negrete, J, Carino-Escobar, RI, Hernández, LG & Falcón, LE 2019, Dendrite Ellipsoidal Neuron Trained by Stochastic Gradient Descent for Motor Imagery Classification. in JA Carrasco-Ochoa, JF Martínez-Trinidad, JA Olvera-López & J Salas (eds), Pattern Recognition - 11th Mexican Conference, MCPR 2019, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11524 LNCS, Springer Verlag, pp. 80-88, 11th Mexican Conference on Pattern Recognition, MCPR 2019, Querétaro, Mexico, 26/06/19. https://doi.org/10.1007/978-3-030-21077-9_8

Dendrite Ellipsoidal Neuron Trained by Stochastic Gradient Descent for Motor Imagery Classification. / Arce, Fernando; Mendoza-Montoya, Omar; Zamora, Erik et al.
Pattern Recognition - 11th Mexican Conference, MCPR 2019, Proceedings. ed. / Jesús Ariel Carrasco-Ochoa; José Francisco Martínez-Trinidad; José Arturo Olvera-López; Joaquín Salas. Springer Verlag, 2019. p. 80-88 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11524 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Dendrite Ellipsoidal Neuron Trained by Stochastic Gradient Descent for Motor Imagery Classification

AU - Arce, Fernando

AU - Mendoza-Montoya, Omar

AU - Zamora, Erik

AU - Antelis, Javier M.

AU - Sossa, Humberto

AU - Cantillo-Negrete, Jessica

AU - Carino-Escobar, Ruben I.

AU - Hernández, Luis G.

AU - Falcón, Luis Eduardo

PY - 2019

Y1 - 2019

N2 - Dendrite ellipsoidal neurons are a novel and different alternative for classification tasks, giving competitive results compared with typical classification methods. Based on k-means++ algorithm, the network allows each dendrite to build a hyperellipsoidal in order to assign each incoming pattern to its respective C class. The main disadvantage of this training algorithm is the lack of accuracy in high dimensional datasets. In this research, we solved this problem by training the dendrite ellipsoidal neuron using stochastic gradient descent. Furthermore, electroencephalography data were acquired during two mental conditions (imaginary movements of the left and right hand) in order to test the new training algorithm. The proposed algorithm outperformed the accuracy acquired by a dendrite ellipsoidal neuron based on k-means++ obtaining 76.02% and 62.77%, respectively. Also, the algorithm was compared with multilayer perceptrons and support vector machines which are some of the most common classifiers used to detect motor-related information in brain signals. These achieved an accuracy of 72.38% and 65.81%, respectively.

AB - Dendrite ellipsoidal neurons are a novel and different alternative for classification tasks, giving competitive results compared with typical classification methods. Based on k-means++ algorithm, the network allows each dendrite to build a hyperellipsoidal in order to assign each incoming pattern to its respective C class. The main disadvantage of this training algorithm is the lack of accuracy in high dimensional datasets. In this research, we solved this problem by training the dendrite ellipsoidal neuron using stochastic gradient descent. Furthermore, electroencephalography data were acquired during two mental conditions (imaginary movements of the left and right hand) in order to test the new training algorithm. The proposed algorithm outperformed the accuracy acquired by a dendrite ellipsoidal neuron based on k-means++ obtaining 76.02% and 62.77%, respectively. Also, the algorithm was compared with multilayer perceptrons and support vector machines which are some of the most common classifiers used to detect motor-related information in brain signals. These achieved an accuracy of 72.38% and 65.81%, respectively.

KW - Dendrite Ellipsoidal Neuron

KW - Electroencephalography

KW - Motor Imagery

KW - Multilayer perceptrons

KW - Stochastic Gradient Descent

KW - Support vector machines

KW - k-means++

UR - http://www.scopus.com/inward/record.url?scp=85068319051&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-21077-9_8

DO - 10.1007/978-3-030-21077-9_8

M3 - Contribución a la conferencia

AN - SCOPUS:85068319051

SN - 9783030210762

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 80

EP - 88

BT - Pattern Recognition - 11th Mexican Conference, MCPR 2019, Proceedings

A2 - Carrasco-Ochoa, Jesús Ariel

A2 - Martínez-Trinidad, José Francisco

A2 - Olvera-López, José Arturo

A2 - Salas, Joaquín

PB - Springer Verlag

T2 - 11th Mexican Conference on Pattern Recognition, MCPR 2019

Y2 - 26 June 2019 through 29 June 2019

ER -

Arce F, Mendoza-Montoya O, Zamora E, Antelis JM, Sossa H, Cantillo-Negrete J et al. Dendrite Ellipsoidal Neuron Trained by Stochastic Gradient Descent for Motor Imagery Classification. In Carrasco-Ochoa JA, Martínez-Trinidad JF, Olvera-López JA, Salas J, editors, Pattern Recognition - 11th Mexican Conference, MCPR 2019, Proceedings. Springer Verlag. 2019. p. 80-88. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-21077-9_8

Dendrite Ellipsoidal Neuron Trained by Stochastic Gradient Descent for Motor Imagery Classification

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this