BCI Competitions III Result
Dataset IVc
ZHOU zongtan, HU Dewen, LIU Yang
Department of Automatic Control,
National University of Defense Technology,
Changsha, Hunan, 410073, P.R.C.
narcz@163.com
Algorithm description:
Preprocessing:
1) Filter signals using a lowpass 25Hz filter and a highpass 2.5Hz filter.
2) For both training and testing data, get data of 1s length after the position of every cue as trials.
3) Compute CSP(Common Spatial Pattern) using training data to reduce the dimensions from 118 channels to 12 channels.
Feature extracting:
1) Compute continuous wavelet transformation(4~24Hz corresponded) of every trials to generate a 2D time-frequency array.
2) Compute a two samples t-Test statistic on every time-frequency point.
3) Construct features using sum of the cwt coefficients weighted by the t-value.
Classifying:
Classify using a linear discriminant analysis(LDA).