Exploiting Spectral and Cepstral Handwriting Features on Diagnosing Parkinson’s Disease

Parkinson’s disease (PD) is the second most frequent neurodegenerative associated with several motor symptoms, including alterations in handwriting, also known as PD dysgraphia. Several computerized decision support systems for dysgraphia have been proposed, however, challenges require new approaches more accurate diagnosis. Therefore, this work adds spectral and cepstral handwriting features to already-used temporal, kinematic statistics features. First, we calculate temporal using displacement; statistic (SF) displacement, horizontal vertical (SDF) (CDF) displacement pressure. Since employed dataset (PaHaW) contains only 37 patients 38 healthy control subjects (HC), then step, augment percentage of smaller training set equal larger. Next, both classes increase patient’s data added random Gaussian noise all augmentations. Third, relevant were selected modified fast correlation-based filtering method (mFCBF). Finally, autoML train test than ten plain ensembled classifiers. Experimental results show that adding kinematics highly improved classification accuracy 98.57%. Our proposed model, lower computational complexities, outperforms conventional state-of-the-art models tasks, which 97.62%.