Feature Extraction and Classification Using Ensemble Method to Enhance Parkinson's Disease Prediction

Parkinson disease (PD), which is more prevalent in persons over 50, has affected millions of people worldwide. The two most common neurodegenerative illnesses are Alzheimer's and Parkinson's, in which majority of people with PD are elderly, since it is a common central nervous system condition. The Clinical diagnosis of the ailment is difficult due to the difficult signs of the sickness. Additionally, it is anticipated to rise over the next ten years, driving up treatment costs. The significance of medical outcomes on patients' financial investments leads to bias, errors, and high clinical expenditures. Precise planning of a treatment is necessary to lower the possibility of a consequence that cannot be retrieved. In this work, the ensemble classifiers are utilized to back up the specialists' diagnosis of PD. Using boosting and bagging algorithm, a collection of classifiers is created, and the classifiers' predictions are then used to categorize new data. The dataset for this study comprises of a variety of biological speech signals from 31 people, including 8 healthy subjects and 23 people with Parkinson's disease. The UCI machine learning (ML) database was used to compile a set of data about Parkinson's disease. The performance metrics used to evaluate the effectiveness of the device include accuracy (ACC), true positive rate (TPR), true negative rate (TNR), positive prediction value (PPV), negative prediction value (NPV), false positive rate (FPR), and false negative rate (FNR). A very high accuracy rate of about 94.4% is indicated by the Parkinson's data.