Title :
MLSP 2010 EEG classification competition and the used method
Author_Institution :
Elektronik ve Haberleş
Abstract :
In the study, the method used in the 6th Annual MLSP Competition (Mind Reading) organized in the 2010 IEEE International Workshop on Machine Learning For Signal Processing (MLSP 2010) is explained. The goal in the competition is to select or design a method (feature extraction, pre-processing and classifier) in order to classify the acquired electroencephalogram (EEG) data from a subject who is looking at the presented satellite images on an LCD monitor that were displayed for specific frequency. For this purpose, a method used in a previous EEG classification competition was adapted to the EEG data of this competition by some modifications. In the competition where the winner was determined as the submission that maximizes the area under the ROC curve, selected method took the second place among 35 participant groups from 14 countries.
Keywords :
"Electroencephalography","Principal component analysis","Signal processing","Machine learning","Art","Brain computer interfaces","Conferences"
Conference_Titel :
Electrical, Electronics and Computer Engineering (ELECO), 2010 National Conference on
Print_ISBN :
978-1-4244-9588-7