Bayesian inference for an adaptive Ordered Probit model: An application to Brain Computer Interfacing

Ji Won Yoon, Stephen J. Roberts, Mathew Dyson, John Q. Gan

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

This paper proposes an algorithm for adaptive, sequential classification in systems with unknown labeling errors, focusing on the biomedical application of Brain Computer Interfacing (BCI). The method is shown to be robust in the presence of label and sensor noise. We focus on the inference and prediction of target labels under a nonlinear and non-Gaussian model. In order to handle missing or erroneous labeling, we model observed labels as a noisy observation of a latent label set with multiple classes (≥2). Whilst this paper focuses on the method's application to BCI systems, the algorithm has the potential to be applied to many application domains in which sequential missing labels are to be imputed in the presence of uncertainty. This dynamic classification algorithm combines an Ordered Probit model and an Extended Kalman Filter (EKF). The EKF estimates the parameters of the Ordered Probit model sequentially with time. We test the performance of the classification approach by processing synthetic datasets and real experimental EEG signals with multiple classes (2, 3 and 4 labels) for a Brain Computer Interfacing (BCI) experiment.

Original languageEnglish
Pages (from-to)726-734
Number of pages9
JournalNeural Networks
Volume24
Issue number7
DOIs
Publication statusPublished - 2011 Sep 1
Externally publishedYes

Fingerprint

Labels
Brain
Computer Systems
Extended Kalman filters
Labeling
Uncertainty
Noise
Electroencephalography
Observation
Computer systems
Sensors
Processing
Experiments
Datasets

Keywords

  • Brain Computer Interfacing
  • Extended Kalman Filter
  • Multi-class classifier
  • Ordered Probit model
  • Sequential decisions

ASJC Scopus subject areas

  • Artificial Intelligence
  • Cognitive Neuroscience

Cite this

Bayesian inference for an adaptive Ordered Probit model : An application to Brain Computer Interfacing. / Yoon, Ji Won; Roberts, Stephen J.; Dyson, Mathew; Gan, John Q.

In: Neural Networks, Vol. 24, No. 7, 01.09.2011, p. 726-734.

Research output: Contribution to journalArticle

Yoon, Ji Won ; Roberts, Stephen J. ; Dyson, Mathew ; Gan, John Q. / Bayesian inference for an adaptive Ordered Probit model : An application to Brain Computer Interfacing. In: Neural Networks. 2011 ; Vol. 24, No. 7. pp. 726-734.
@article{fada16e4fc904fa58eb3ef075313d211,
title = "Bayesian inference for an adaptive Ordered Probit model: An application to Brain Computer Interfacing",
abstract = "This paper proposes an algorithm for adaptive, sequential classification in systems with unknown labeling errors, focusing on the biomedical application of Brain Computer Interfacing (BCI). The method is shown to be robust in the presence of label and sensor noise. We focus on the inference and prediction of target labels under a nonlinear and non-Gaussian model. In order to handle missing or erroneous labeling, we model observed labels as a noisy observation of a latent label set with multiple classes (≥2). Whilst this paper focuses on the method's application to BCI systems, the algorithm has the potential to be applied to many application domains in which sequential missing labels are to be imputed in the presence of uncertainty. This dynamic classification algorithm combines an Ordered Probit model and an Extended Kalman Filter (EKF). The EKF estimates the parameters of the Ordered Probit model sequentially with time. We test the performance of the classification approach by processing synthetic datasets and real experimental EEG signals with multiple classes (2, 3 and 4 labels) for a Brain Computer Interfacing (BCI) experiment.",
keywords = "Brain Computer Interfacing, Extended Kalman Filter, Multi-class classifier, Ordered Probit model, Sequential decisions",
author = "Yoon, {Ji Won} and Roberts, {Stephen J.} and Mathew Dyson and Gan, {John Q.}",
year = "2011",
month = "9",
day = "1",
doi = "10.1016/j.neunet.2011.03.019",
language = "English",
volume = "24",
pages = "726--734",
journal = "Neural Networks",
issn = "0893-6080",
publisher = "Elsevier Limited",
number = "7",

}

TY - JOUR

T1 - Bayesian inference for an adaptive Ordered Probit model

T2 - An application to Brain Computer Interfacing

AU - Yoon, Ji Won

AU - Roberts, Stephen J.

AU - Dyson, Mathew

AU - Gan, John Q.

PY - 2011/9/1

Y1 - 2011/9/1

N2 - This paper proposes an algorithm for adaptive, sequential classification in systems with unknown labeling errors, focusing on the biomedical application of Brain Computer Interfacing (BCI). The method is shown to be robust in the presence of label and sensor noise. We focus on the inference and prediction of target labels under a nonlinear and non-Gaussian model. In order to handle missing or erroneous labeling, we model observed labels as a noisy observation of a latent label set with multiple classes (≥2). Whilst this paper focuses on the method's application to BCI systems, the algorithm has the potential to be applied to many application domains in which sequential missing labels are to be imputed in the presence of uncertainty. This dynamic classification algorithm combines an Ordered Probit model and an Extended Kalman Filter (EKF). The EKF estimates the parameters of the Ordered Probit model sequentially with time. We test the performance of the classification approach by processing synthetic datasets and real experimental EEG signals with multiple classes (2, 3 and 4 labels) for a Brain Computer Interfacing (BCI) experiment.

AB - This paper proposes an algorithm for adaptive, sequential classification in systems with unknown labeling errors, focusing on the biomedical application of Brain Computer Interfacing (BCI). The method is shown to be robust in the presence of label and sensor noise. We focus on the inference and prediction of target labels under a nonlinear and non-Gaussian model. In order to handle missing or erroneous labeling, we model observed labels as a noisy observation of a latent label set with multiple classes (≥2). Whilst this paper focuses on the method's application to BCI systems, the algorithm has the potential to be applied to many application domains in which sequential missing labels are to be imputed in the presence of uncertainty. This dynamic classification algorithm combines an Ordered Probit model and an Extended Kalman Filter (EKF). The EKF estimates the parameters of the Ordered Probit model sequentially with time. We test the performance of the classification approach by processing synthetic datasets and real experimental EEG signals with multiple classes (2, 3 and 4 labels) for a Brain Computer Interfacing (BCI) experiment.

KW - Brain Computer Interfacing

KW - Extended Kalman Filter

KW - Multi-class classifier

KW - Ordered Probit model

KW - Sequential decisions

UR - http://www.scopus.com/inward/record.url?scp=79960890851&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79960890851&partnerID=8YFLogxK

U2 - 10.1016/j.neunet.2011.03.019

DO - 10.1016/j.neunet.2011.03.019

M3 - Article

C2 - 21493037

AN - SCOPUS:79960890851

VL - 24

SP - 726

EP - 734

JO - Neural Networks

JF - Neural Networks

SN - 0893-6080

IS - 7

ER -