Classification of Upper Limb Movement Imaginations Based On a Hybrid Method of Wavelet Transform and Principal Component Analysis for Brain-Computer Interface Applications

The Brain-Computer Interface in the last decade, the scientific journey has received increasing attention, and the holding of several international competitions and scientific challenges around the world is proof of this claim. In this paper, a six-step algorithm is used to classify the perceptions of limb movements. In the first step, a collection of 288 electroencephalogram data was collected from the BCI Competition Database of 2005. In the second step, data noise reduction was performed using a wavelet bank filter. In the third step, the meow and beta rhythms of the signal in the central region were extracted using a wavelet frequency domain time domain display. In the fourth step, a set of temporal, frequency, and nonlinear properties were extracted from each sub-band, and in the fifth step, the feature space was reduced using principal component analysis. In the sixth step, the feature set was considered as the input of the two nearest neighbor classifiers, the backup vector machine, and the decision tree. All simulations have been executed and implemented under MATLAB software. The results show that the support vector machine classifier with nonlinear kernel and nearest neighbor classifier has an efficiency of more than 80%.

کلیدواژه ها:

Brain and computer interface ، electroencephalogram ، Mayo rhythm ، beta rhythm ، wavelet transform ، واسط مغز و کامپیوتر ، الکتروانسفالوگرام ، ریتم میو ، ریتم بتا ، تبدیل موجک