<![CDATA[This study utilized a 20-sample dataset with 23 acoustic attributes reflecting the patient's vocal characteristics, including parameters such as MDVP:Fo(Hz), MDVP:Fhi(Hz), MDVP:Flo(Hz), jitter, shimmer, and harmonic values. The target attribute used was status, which indicates whether the patient is indicated as having Parkinson's (1) or not (0). The study aimed to build a data mining-based prediction model using RapidMiner to support an information system for early detection of Parkinson's Disease. The analysis stages followed the CRISP-DM framework, which includes data understanding, preprocessing, modeling, and evaluation. The algorithms tested included Support Vector Machine, Random Forest, and Naïve Bayes using a cross-validation scheme. The experimental results showed that Random Forest provided the most consistent performance on this small dataset with more stable accuracy and prediction confidence values compared to other algorithms. These findings confirm that acoustic features of voice can serve as effective early indicators for detecting Parkinson's, while strengthening RapidMiner's role as an efficient analytical platform in data-driven medical research.]]>
Copyrights © 2026
