Extraction des caractéristiques pour la classification de la maladie de Parkinson - Archive ouverte HAL Access content directly
Reports Year :

Extraction des caractéristiques pour la classification de la maladie de Parkinson

Abstract

Parkinson's disease (PD) is a chronic, progressive disorder of movement, which means that the symptoms persist and worsen over time. While the voices seem to be affected as well as the movements of the disease, the development of a simple test that relies on voice by analyzing a number of parameters that characterizes it, and on the basis of registration of classifying the patient following two "parkinsonism or healthy" situations can be a track that appears increasingly to be very promising. This led us to develop a system diagnostic aid for the early detection of the disease so we use the medical data base created by Max little for implantation of the neural network classifier and we compared the results obtained with the neural classifier while applying the results of step reduction parameters in which we adopted independent component analysis "ICA". The ability of neural network classification has highlighted its effectiveness for faster learning with reduced structure, through the optimization method or reductions "ICA"
Fichier principal
Vignette du fichier
RapportPARKINSON.pdf (333.33 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-00846805 , version 1 (20-07-2013)

Identifiers

  • HAL Id : hal-00846805 , version 1

Cite

Soumia Benikhlef, Bendimerad El Batoul, Nesma Settouti. Extraction des caractéristiques pour la classification de la maladie de Parkinson. 2013. ⟨hal-00846805⟩

Collections

LARA
469 View
1545 Download

Share

Gmail Facebook Twitter LinkedIn More