Identification of Suitable Discrete Wavelet Order for Motor Imagery and Motor Movement Waveforms

The goal of this paper was to determine the best wavelet type for analyzing the datasets for Motor Imaginary and Motor Movement. This was done by adding an additional set of Discrete Wavelet Transform features to the existing set of time-domain and frequency-domain features calculated. The data was then classified using a multi-modal classification model whose input features were the numerical values and an output feature. The effectiveness of Coiflet’s, Daubechies, Bior, and Symlet’s families of wavelets in improving the results was validated by calculating various performance metrics after training with a multi-modal classification model. The findings obtained, when coupled with a Brain–Computer Interface, could potentially be applied in the development of a device capable of replicating these hand movements, such as a robotic arm.