Brain-computer interfaces (BCIs) aim to provide a non-muscular channel to communicate with the external world through the use of the brain Electroencephalograph (EEG) activity. A crucial step in such an operation is brain signal processing methods. BCI systems use EEG as it is practical, noninvasive, cheap and has real time capability imaging technology. BCI's efficiency is dependent on brain signal processing methods which classify brain signal patterns accurately in various tasks. The presence of artifacts in raw EEG signal makes it necessary to preprocess the signal for feature extraction. This paper presents a BCI system preprocessing and extracting features from EEG signals through the use of Walsh Hadamard Transform (WHT). Signal classification is done using Bagging techniques.