Kumar Ajay, Singh Jay Parkash, Paygude Priyanka, Daimary Rachan, Prasad Sandeep
Department of Computer Science & Engineering, Manipal University Jaipur, Rajasthan, India.
Department of Information Technology, Bharati Vidyapeeth (Deemed to be University) College of Engineering Pune, Maharashtra, India.
Digit Health. 2025 Jun 6;11:20552076251342878. doi: 10.1177/20552076251342878. eCollection 2025 Jan-Dec.
To investigate the potential of voice analysis-specifically sustained vowel phonation-as a non-invasive, cost-effective diagnostic method for early detection of Parkinson's disease (PD) using machine learning techniques.
A publicly available dataset from the University of California, Irvine (UCI) repository, comprising 252 voice recordings (188 from PD patients and 64 from healthy individuals), was analyzed. Machine learning classifiers, including k-nearest neighbors (KNN), AdaBoost, and artificial neural networks (ANNs), were trained and tested on the dataset. Model evaluation was conducted using accuracy, precision, recall, F1 score, and area under the receiver operating characteristic curve. Kernel density estimation was applied to visualize and interpret classifier performance.
Among the classifiers, KNN demonstrated the best performance with an accuracy of 98.52% and a mean accuracy of 97.33%. ANN and AdaBoost achieved mean accuracies of 93.15% and 91.77%, respectively. All models performed well across standard evaluation metrics, indicating strong discriminative ability for detecting PD from voice data.
The study confirms the feasibility of using sustained vowel phonation and machine learning for early PD diagnosis. The KNN classifier, in particular, shows excellent diagnostic accuracy. These findings support the integration of voice-based machine learning tools into clinical workflows, potentially enhancing early detection and management of PD.
利用机器学习技术，研究语音分析——特别是持续元音发声——作为一种用于帕金森病（PD）早期检测的非侵入性、经济高效的诊断方法的潜力。
分析了来自加利福尼亚大学欧文分校（UCI）存储库的一个公开可用数据集，该数据集包含252个语音记录（188个来自PD患者，64个来自健康个体）。在该数据集上训练和测试了包括k近邻（KNN）、AdaBoost和人工神经网络（ANN）在内的机器学习分类器。使用准确率、精确率、召回率、F1分数和受试者工作特征曲线下面积进行模型评估。应用核密度估计来可视化和解释分类器性能。
在分类器中，KNN表现最佳，准确率为98.52%，平均准确率为97.33%。ANN和AdaBoost的平均准确率分别为93.15%和91.77%。所有模型在标准评估指标上表现良好，表明从语音数据中检测PD具有很强的判别能力。
该研究证实了使用持续元音发声和机器学习进行PD早期诊断的可行性。特别是KNN分类器，显示出优异的诊断准确性。这些发现支持将基于语音的机器学习工具整合到临床工作流程中，可能会加强PD的早期检测和管理。
