Author(s): Nooritawati Md. Tahir | Hany Hazfiza Manap
Journal: Journal of Applied Sciences
ISSN 1812-5654
Volume: 12;
Issue: 2;
Start page: 180;
Date: 2012;
VIEW PDF
DOWNLOAD PDF 
Original page
Keywords: support vector machine | Parkinson disease | gait recognition | artificial neural network
ABSTRACT
This study discussed the ability of two machine classifiers namely Artificial Neural Network (ANN) and Support Vector Machine (SVM) in distinguishing gait pattern during self-selected speed walking due to the effect of motor Parkinson Disease (PD). There are three gait parameters that is utilized as features in classifying PD gait and normal subjects namely basic spatiotemporal, kinematic and kinetic. Firstly, the input features are pre-processed using two types of normalization technique specifically intra group as well as inter group normalization. Additionally, all the three features are classified solely followed by implementation of data fusion. Then, the effectiveness of the features vectors to identify PD patients or vice versa is evaluated based as inputs to both classifiers. Initial findings showed that basic spatiotemporal solely as feature vectors based on intra group normalization technique contributed perfect classification for both ANN and SVM as classifiers.
Journal: Journal of Applied Sciences
ISSN 1812-5654
Volume: 12;
Issue: 2;
Start page: 180;
Date: 2012;
VIEW PDF
DOWNLOAD PDF Original pageKeywords: support vector machine | Parkinson disease | gait recognition | artificial neural network
ABSTRACT
This study discussed the ability of two machine classifiers namely Artificial Neural Network (ANN) and Support Vector Machine (SVM) in distinguishing gait pattern during self-selected speed walking due to the effect of motor Parkinson Disease (PD). There are three gait parameters that is utilized as features in classifying PD gait and normal subjects namely basic spatiotemporal, kinematic and kinetic. Firstly, the input features are pre-processed using two types of normalization technique specifically intra group as well as inter group normalization. Additionally, all the three features are classified solely followed by implementation of data fusion. Then, the effectiveness of the features vectors to identify PD patients or vice versa is evaluated based as inputs to both classifiers. Initial findings showed that basic spatiotemporal solely as feature vectors based on intra group normalization technique contributed perfect classification for both ANN and SVM as classifiers.
