Study On The Classification Of Two Types Of EEG

With the rapid development of computer technology and application of the signal processing, it is expected to become possible for the brain-computer int-erface (BCI) system going from the laboratory to the real world. The BCI system is a direct exchange of information and control channel between brain and com-puters or other electronic devices. It is a new information exchange system, whi-ch is not dependent on conventional brain output pathways (peripheral nerve and muscle tissue). The new exchange of information and control technology will be able to provide a communication with the outside world and control the new ch-annels for paralyzed patients, especially those who lost their basic limb motor f-unction, but thinking is normal, which is obtaining increasingly and more attent-ion.Now, for the limit of the accuracy and speed of the classification on the EEG, the BCI system has not been brought from the laboratory into the real world. Th-e BCI experiment studied in this paper uses the BCI2003 competition data to cl-assify the EEG. In this paper, we propose a concept of EEG trend, and use supp-ort vector machine (SVM) as classification algorithms. First, the multi-channel EEG data pass the low-pass filter and the band pass filter, respectively. Second, use the time window to filter them form time domain, select the section of the most significant phenomenon of performance data, and extract features through the common spatial subspace decomposition (CSSD) from the signals. At last, based on the extracted features, classify them through the SVM training. The r-ecognition rate is 85%-95%, the accuracy for the EEG classification increases a lot.The classification algorithm used in paper is a new algorithm, which is im-proved basing on the works of previous researchers. It has been proved that the algorithm has higher recognition rate, which provides new ideas and new meth-ods for the classification of EEG.