"Vavre\u010Dka, Michal" . "RIV/68407700:21230/13:00212220" . . . . "2013-07-19+02:00"^^ . "EEG Feature Selection Based on Time Series Classification"@en . . . "EEG; Reference Frames; Spatial Navigation; Virtual Environment"@en . . "http://link.springer.com/chapter/10.1007%2F978-3-642-39712-7_40" . "EEG Feature Selection Based on Time Series Classification" . . "2"^^ . "10.1007/978-3-642-39712-7_40" . . "RIV/68407700:21230/13:00212220!RIV14-MSM-21230___" . "978-3-642-39711-0" . "New York" . "Lecture Notes in Computer Science" . "2"^^ . "[99D995452BE1]" . . . . . . "Lhotsk\u00E1, Lenka" . . . "8"^^ . "71281" . "21230" . "We propose novel method of EEG signal analysis based on classification of feature time series. The algorithm classifies sequences of feature values and it calculates the error rate both for each time step and overall sequence. We compared the performance of the algorithm with a standard feature selection method based on forward inter-intra criterion. Both algorithms selected similar features. The algorithm was tested on the EEG data from 2 experiments focused on of spatial navigation and orientation. Participants traversed through the virtual tunnels and they could adopt two different reference frames (allocenctric and egocentric) to solve the task. The EEG signal was recorded within both tasks and the methods of feature extraction and both standard and timeseries selection and classification were applied to it. We identified differences between the groups of participants adopting allocentric and egocentric frames of reference in the parietal and central electrodes in right hemisphere. The novel algorithm provided more detail analysis of the EEG features compared to classic feature classification." . "Springer-Verlag" . . "0302-9743" . . "P(EE2.3.30.0049), P(GPP407/11/P696), Z(MSM6840770012)" . . . "EEG Feature Selection Based on Time Series Classification"@en . "We propose novel method of EEG signal analysis based on classification of feature time series. The algorithm classifies sequences of feature values and it calculates the error rate both for each time step and overall sequence. We compared the performance of the algorithm with a standard feature selection method based on forward inter-intra criterion. Both algorithms selected similar features. The algorithm was tested on the EEG data from 2 experiments focused on of spatial navigation and orientation. Participants traversed through the virtual tunnels and they could adopt two different reference frames (allocenctric and egocentric) to solve the task. The EEG signal was recorded within both tasks and the methods of feature extraction and both standard and timeseries selection and classification were applied to it. We identified differences between the groups of participants adopting allocentric and egocentric frames of reference in the parietal and central electrodes in right hemisphere. The novel algorithm provided more detail analysis of the EEG features compared to classic feature classification."@en . . . "Heidelberg" . . "EEG Feature Selection Based on Time Series Classification" .