Parkinson's Disease Detection Using Data Mining Models: A Comparative Study

Data Mining techniques have emerged in the healthcare industry over the past years in determining the occurrences and development of diseases. This paper aim is to conduct a comparative study between three Data Mining models in predicting Parkinson’s Disease using biological voice measurements dataset. The three models that are evaluated are: Random Forest (RF), K-Nearest Neighbors (KNN) and Support Vector Machine (SVM). Data oversampling using Synthetic Minority Oversampling Technique (SMOTE) method was used to balance the data and feature selection was applied to select the most relevant attributes. The CRISP-DM framework is used to perform the analysis using Weka machine learning tool. The results showed that the KNN model obtained the best accuracy of 94.11% on balanced dataset.