Date of Publication :18th October 2023

Abstract: Abstract— Parkinson’s disease (PD) is a progressive disease leading to muscle movement disorder, which is often difficult to diagnose, especially in the early stages. Therefore, it furtively progresses throughout the body, rendering its treatment and control impossible. Currently, data mining science allows for making reliable decisions at the best possible time through the collection and comprehensive analysis of key information. This study aims to provide a simple and cost-effective method for early diagnosis of PD and to select appropriate features. This research reviews a model based on ontology and machine learning (ML) methods for PD diagnosis. This model examines voice features extracted during ontology matching using three classifiers: decision tree, k-nearest neighbors (kNN), and support vector machine (SVM). Finally, the best classifier is introduced as the final result based on the results obtained for this database. Among all evaluated indicators, the most important was attached to age and shimmer (amplitude variation of the sound) at the same time interval for PD diagnosis. The 4.5C classifier dataset obtained the best result for PD diagnosis with 93.2% accuracy, which is highly satisfactory because a low-cost method was employed to develop voice features.

Reference :