Data Segments with Different Wavelet Bands and Stages of Voting for the Discrimination of Parkinson Tremor from Essential Tremor Using Accelerometer and EMG Signals
Zaynab Riyadh K. Al-Hakim
Department of Computer Engineering, College of IT, University of Bahrain
Abstract—A new idea for the identification of Parkinson tremor from essential tremor is presented in this paper. Segments of data of accelerometer and surface EMG signals are used with different wavelet bands for the idea of discrimination of Parkinson tremor from essential tremor. The data used are from the University of Kiel, Germany. The data are 41 training subjects: 21 with Essential-tremor (ET) and 19 with Parkinson-disease (PD). Another 40 subjects of test data have 20 PD and 20 ET subjects, are used to test the technique. In this study three different data segments, each with its best fit wavelet band for each signal are selected. Then, a two-stages voting between the results is obtained. The discrimination efficiency on test data resulted 100% sensitivity, 85% specificity and 92.5% accuracy.
Index Terms—Wavelet-band, Data-Segment, Voting, Parkinson Tremor, Essential Tremor, EMG, Accelerometer Signals, Discrimination
Cite: Zaynab Riyadh K. Al-Hakim, "Data Segments with Different Wavelet Bands and Stages of Voting for the Discrimination of Parkinson Tremor from Essential Tremor Using Accelerometer and EMG Signals," Journal of Medical and Bioengineering, Vol. 3, No. 2, pp. 128-132, June 2014. Doi: 10.12720/jomb.3.2.128-132
Index Terms—Wavelet-band, Data-Segment, Voting, Parkinson Tremor, Essential Tremor, EMG, Accelerometer Signals, Discrimination
Cite: Zaynab Riyadh K. Al-Hakim, "Data Segments with Different Wavelet Bands and Stages of Voting for the Discrimination of Parkinson Tremor from Essential Tremor Using Accelerometer and EMG Signals," Journal of Medical and Bioengineering, Vol. 3, No. 2, pp. 128-132, June 2014. Doi: 10.12720/jomb.3.2.128-132
Array