Title: Unsupervised EEG biosignal discrimination

Authors: Olga Georgieva; Sergey Milanov; Petia Georgieva

Addresses: Faculty of Mathematics and Informatics, Sofia University 'St. Kl. Ohridski', Sofia, Bulgaria ' Faculty of Mathematics and Informatics, Sofia University 'St. Kl. Ohridski', Sofia, Bulgaria ' Signal Processing Lab, IEETA, University of Aveiro, Aveiro, Portugal

Abstract: This paper assesses the ability of unsupervised learning to discriminate positive and negative human emotions evoked while the person is exposed to images with high emotional content. The investigation is focused in two different settings - multi-channel EEG data from individual subjects (person data) and multiple subjects EEG data from individual channels (channel data). The person data is used for intra-subject clustering and the channel data is used for inter-subject clustering. Attribute selection procedure is first applied in order to select the most important data features. Secondly, K-means and fuzzy-C-means clustering techniques are comparatively applied in order to extract specific knowledge of the existing dependences and thus decode the two emotional states.

Keywords: unsupervised learning; cluster analysis; biosignal discrimination; EEG data; electroencephalographs; human emotions; image content; emotional content; clustering.

DOI: 10.1504/IJRIS.2014.066249

International Journal of Reasoning-based Intelligent Systems, 2014 Vol.6 No.3/4, pp.118 - 125

Received: 08 May 2021
Accepted: 12 May 2021

Published online: 09 Dec 2014 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article