Performance Analysis of Machine Learning Classification for Parkinson's Disease from Biomedical Audio Data

• Published in: 2023 Intelligent Computing and Control for Engineering and Business Systems (ICCEBS) pp. 1 - 5
• Main Authors: C, Santhosh Kumar, A, Suagnthi, N, Sarmiladevi, A, Indhuja, G, Venkatesan
• Format: Conference Proceeding
• Language: English
• Published: IEEE 14.12.2023
• Subjects: Classification; Classification algorithms; Genetics; Imaging; Logistic regression; Machine Learning; Machine learning algorithms; Parkinson's disease; Sensitivity
• DOI: 10.1109/ICCEBS58601.2023.10448605

Millions of people throughout the world suffer from the crippling neurological ailment known as Parkinson's disease (PD). While there is no cure for PD, early recognition and intervention can improve a patient's quality of life. The use of machine learning for the categorization of Parkinson's disease using biomedical audio data has gained popularity in recent years. In this study, we employ classification algorithms to categorize Parkinson's disease patients using a clinical, genetic, and audio dataset. In this paper, we analyze the effectiveness of various classification algorithms for detecting Parkinson's disease, such as Logistic Regression, k-NN, decision tree, random forest, and an ensemble model employing AdaBoost classifier. Our results demonstrate that the accuracy of the decision tree classifier algorithm outperforms the other algorithms by 92.59% on our dataset. In particular, the study highlights the potential use of algorithms for machine learning in identifying and diagnosing the early stages of Parkinson's disease.