Machine Learning Algorithms to Diagnose Parkinson’s Disease Using Vocal Data

Parkinson’s disease is a kind of chronic ailment and the suffering from this chronic ailment has a neurodegenerative disorder. With time they gradually deteriorate both their motor and non-motor abilities. It is challenging to identify Parkinson’s disease (PD) as its symptoms are nearly the same as normal aging and movement disorders. Early PD diagnosis is the preferred diagnosis for the effective treatment and potential recovery of PD patients. The study investigates deployment of multiple machine learning (ML) models for the effective diagnosis of PD’s patients. Here, a detailed comparative analysis of different ML algorithms such as supervised machine learning algorithms as well as boosting algorithms is portrayed. These algorithms include logistic regression (LR), support vector machine (SVM), meta classifier B (MC-B), random forest (RF), and extreme gradient boost (XGBoost) for the performance evaluation of each classifier. Python platform is employed for the simulation study, and the findings show that XGBoost outperforms other classifiers in terms of greatest recall of 1.00, precision of 0.87, and prediction accuracy of 90%.