Process used for prediction of Labels for BCI Competition 2003. 

i. The amount of data points in the training and test set were reduced
through decimation.

ii. Features from each channel of the data sets were obtained using
discriminant pursuit, each channel being analyzed separately.

iii. The predicted labels for the test set were obtained by using
Support Vector Machine analysis.

Developed by Juma Mbwana, Yale University
Assisted by Mark Laubach, Yale University School of Medicine