Parkinson's disease diagnosis using modular systems

Paper i proceeding, 2013

In this paper, we present two modular systems for Parkinson's disease diagnosis. Also, we compare the frequency and chaotic behavior of rest tremor velocity in the index finger of some parkinsonian and healthy subjects. The proposed methods consist of two different modules, first, high-dimensional features are compressed by local linear and nonlinear principal component analysis (PCA) techniques and then, the features are classified by neural classifiers. The results indicate the efficiency of modular systems in Parkinson's disease diagnosis.