Recognition of Parkinson's ailment by using various machine learning procedures

Parkinson's disease (PD) is a progressive and degenerative neurological disorder characterized by reduced dopamine levels in the central nervous system. It ranks as the second most prevalent neurodegenerative disease worldwide. Diagnosing PD remains challenging and time-consuming, as there is no definitive protocol for its confirmation. The concept of active aging aims to optimize various aspects of fitness during the aging process, ultimately enhancing the quality of life. However, most initiatives have primarily focused on normal aging, with limited attention given to elderly individuals facing chronic conditions like PD. The method employs machine learning classifiers and features derived from metrics of fluctuation magnitude and fluctuation dynamics, obtained through detrended fluctuation analysis. By leveraging these techniques, the proposed method offers a potential advancement in the diagnosis of Parkinson's disease. Based on the accuracies of six classification methods, including support vector machine (SVM), K-nearest neighbor (KNN), random forest (RF), logistic regression (LR), decision tree (DT) and neural network (NN), the assessment and comparisons are performed. The Parkinson's dataset was subjected to analysis using various algorithms, employing a tenfold cross-validation approach. For training and testing, the dataset was split into 60% for training and 40% for testing in each fold. After conducting the evaluations, the results indicated that the neural network algorithm outperformed the other models with an impressive test accuracy of 98.12%.