Computational model of affective states profiling using commercial 14-channel wireless EEG

Yaacob, Hamwira Sakti and Abdul Rahman, Abdul Wahab (2015) Computational model of affective states profiling using commercial 14-channel wireless EEG. In: 28th International Conference on Computer Applications in Industry and Engineering, 12-14 October 2015, San Diego, California, USA.

	PDF (Computational Model of Affective States Profiling using Commercial 14-channel Wireless EEG) - Submitted Version Restricted to Repository staff only Download (1MB) | Request a copy
	PDF (Computational Model of Affective States Profiling using Commercial 14-channel Wireless EEG) - Supplemental Material Restricted to Repository staff only Download (173kB) | Request a copy

Abstract

Many studies apply computational models for affective states profiling through brain activities manifestations which are captured using electroencephalogram (EEG) devices. Although various kinds of devices are available for research purposes, such products are not available off-the-shelf. Moreover, most of EEG devices refrain users from certain movements. This becomes a challenge for capturing EEG signals during active and vibrant tasks. Thus, the aim of this study is to explore the potential of using commercial 14-channel wireless EEG for capturing the brain signals and affective states profiling through computational approach. In this study, power spectral density (PSD) is used as features. Different approaches of feature extractions are compared including the average performance of affective state classification using exclusively alpha and beta frequency bands, the average of energy density over alpha through beta frequency bands and the combination of alpha and beta power spectral density. Multilayer perceptron (MLP) neural networks are used for classification of affective states based on valence and arousal. Based on 11-fold cross validation, the classification of affective states using spectral features containing alpha and beta frequency bands produced accuracy above 80 %. In short, results have shown that 14-channel EPOC neuroheadset is viable for performing affective states profiling.

Item Type:	Conference or Workshop Item (Plenary Papers)
Additional Information:	4870/48693
Uncontrolled Keywords:	Affective, valence, arousal, EEG, EPOC
Subjects:	B Philosophy. Psychology. Religion > BF Psychology > BF511 Affection. Feeling. Emotion Q Science > QA Mathematics > QA76 Computer software T Technology > T Technology (General)
Kulliyyahs/Centres/Divisions/Institutes (Can select more than one option. Press CONTROL button):	Kulliyyah of Information and Communication Technology > Department of Computer Science Kulliyyah of Information and Communication Technology > Department of Computer Science
Depositing User:	Dr Hamwira Yaacob
Date Deposited:	20 Jan 2016 11:18
Last Modified:	24 May 2016 09:51
URI:	http://irep.iium.edu.my/id/eprint/48693

Actions (login required)

View Item