The Electroencephalogram (EEG) of a human brain constitutes the rudimentary aspect of determination in cognizing the various neurological and corporeal variations that an individual may be diagnosed with. The neoteric times have evinced the EEG metrics to be quintessential in narrowing the health detriments in the human body. The processing of brain signals has entailed various unconventional technologies, and their meticulous understanding forms to be a crucial dimensionality of medical studies. Research apropos to identifying bradykinesia constitutes largely on the overall observation of the clinical parameters such as neurological fluctuations and motor volatilities. This Indagation pivots on segregating the alpha signal waves from the brain signals through the Discrete Wavelet transform level-5-Daubechies-2 into different phases such as relaxed, active and sleeping mode. The proposed algorithm pivots on enhancing the classification accuracy of identifying bradykinesia through the various modes by effectuating the Multilevel Support Vector Machine (SVM) with the Radial-basis function (RBF) kernel and entropy bandwidth frequency calculation using the Thomson-Welch Multitaper power spectral density estimation method. The results are successfully obtained, and the simulations are implemented in MATLAB.