Complex Systems

Daubechies Wavelet Cepstral Coefficients for Parkinson’s Disease Detection Download PDF

Soumaya Zayrit
Taoufiq Belhoussine Drissi
Abdelkrim Ammoumou

Laboratory Industrial Engineering, Information Processing and Logistics
University Hassan II, Faculty of Science Ain Chok, Casablanca, Morocco

Benayad Nsiri
Research Center STIS, M2CS Mohammed V University in Rabat
Higher School of Technical Education of Rabat (ENSET), Morocco


The aim of this paper is to evaluate the performance of the approach that focuses on support vector machine (SVM) classification of vocal recording to differentiate between patients affected by Parkinson’s disease (PD) and healthy patients. Our study was based on the condition of 38 patients, some of whom are healthy and others who suffer from PD. The study was carried out as follows: The extraction of cepstral coefficients was reached through the transformation of the speech signal by discrete wavelet transform (DWT) and also through cepstral analysis by using the mel scale. At the end, a classification was done by the use of the two kernels linear and radial basis function (RBF) of the SVM classifier.

Keywords: Parkinson’s disease; Daubechies wavelet; MFCC; SVM

Cite this publication as:
S. Zayrit, T. Belhoussine Drissi, A. Ammoumou and B. Nsiri, “Daubechies Wavelet Cepstral Coefficients for Parkinson’s Disease Detection,” Complex Systems, 29(3), 2020 pp. 729–739.