Research Area:  Machine Learning

Abstract:

Parkinsons disease (PD) is a prevalent neurodegenerative disorder that has prompted the development of telediagnosis and remote monitoring systems. Dysphonia, a common symptom in the early stages of PD, affects approximately 90% of patients. Therefore, testing for persistent pronunciation or dysphonia in continuous speech can aid in the diagnosis of PD. Our study utilized speech signals from 252 subjects as the dataset. In this study, language signal features were used as input to machine learning algorithms, and the resulting classifiers were integrated to improve accuracy in the classification of Parkinsons disease (PD). The experimental results demonstrated a diagnostic accuracy of up to 95% using these machine learning algorithms. Additionally, a method of feature extraction based on clinical experience was presented for analyzing subjects language signals.

Keywords:  

Author(s) Name:  Linlin Yuan, Yao Liu, Hsuan-Ming Feng

Publisher name:  Springer

DOI:  10.1007/s11761-023-00372-w

Volume Information:  Volume 18, pages 101-107, (2024)

Paper Link:   https://link.springer.com/article/10.1007/s11761-023-00372-w