Publication Date: 2024/01/09
Abstract: This paper investigates using machine learning to help diagnose Parkinson's disease early on. It specifically looks at two aspects - hand movements and vocal features. Unique datasets for both the symptoms are explored. Specialized techniques extract the most distinguishing hand motions and speech characteristics as biomarkers. Unlike traditional methods relying on one feature only, this multimodal approach combines both hand movement and voice biomarkers into one computational model. Overall, the study illustrates promise for machine learning tools enabling earlier intervention for medical purposes, not individual diagnosis. The focus remains on aiding clinicians rather than replacing specialized assessments.
Keywords: Deep Learning, Feature Extraction, Neural Networks, Pattern Recognition, Multimodal Approach, Computational Models.
DOI: https://doi.org/10.5281/zenodo.10474047
PDF: https://ijirst.demo4.arinfotech.co/assets/upload/files/IJISRT23DEC1898.pdf
REFERENCES