Hypokinetic Rigid Syndrome: Early Prediction of Parkinson’s Disease using Ensemble classifiers

Parkinson’s disease (PD) is a neurodegenerative illness that manifests as both motor and non-motor symptoms. About 90% of patients experience early onset speech problems. This paper proposes a machine learning system to help with early diagnosis and maybe impede the progression of disease by using ensemble classifiers on speech feature datasets from the UCI repository. There are three instances of speech characteristics in each of the numerous feature subsets from 252 participants that make up the dataset. Highly correlated features are integrated using Principal Component Analysis (PCA) and Linear Discriminate Analysis (LDA) in order to preserve information integrity. Ensemble classifiers such as bagging, Extreme Gradient Boosting (XgBoost), Gradient Boosting Machine (GBM), and Adaptive Boosting (AdaBoost) are used for PD prediction. Through the use of methods like grid search and random search, hyperparameter tuning is used to maximize model performance. Accuracy, Precision, Recall, F1 score, and Support are some of the evaluation metrics. The experimental results show that boosting classifiers achieve more accuracy than bagging classifiers, grid search improves outcomes over random search, and PCA outperforms LDA.