Parkinson Disease Prediction and Classification Using Ensemble Stacking Learning Algorithm

Abstract

Data Mining (DM) is a pragmatic method to find patterns in massive datasets, representing knowledge that is implicitly stored and to focus on problems of feasibility, utility, efficacy, and scalability. Medical data mining is a huge field of study, where techniques are used to address issues on diagnosis, treatment, and to explore the knowledge of disease identification. A neurological disease known as Parkinson Disease (PD) has reportedly affected the majority of people across the world. According to recent studies, voice defects are identified in 90% of PD patients. PD is a disorder of the central sensory system in human body that affects movement and causes tremors. Unfortunately, it can be difficult to identify PD in its early stages. A number of measures available to identify Parkinson Disease. As a result, voice estimations can be utilized to identify the state of impacted persons. This research work proposed a novel method called as Ensemble Stacking Learning Algorithm (ESLA), a classifier for identifying PD on the collected data. To calculate the performance, the proposed ensemble method is compared with other existing technique and shows the improved classification ability of this proposed method. It is exhibited that the proposed methods for PD patients, which creates the most reliable outcomes and accomplishes the highest accuracy. This ensemble approach uses various existing classifiers like Random Forest (RF), XGBoost (Extreme Gradient Boost), Linear Regression, Ada Boost (AB), and Multi-Layer Perceptron (MLP) and compared results with the proposed method for a better prediction of accuracy. Finally, performance of the proposed method among the chosen algorithms is suggested for the prediction of disease and gives future directions.