Parkinson's Disease (PD) affects speech in measurable ways, yet most existing models rely on numerical features, limiting interpretability and adaptability for language-based analysis. To address this, we convert structured speech features into clinically meaningful textual symptom descriptions and leverage Sentence-BERT (S-BERT) for PD classification. To improve model robustness and diversity, we apply textual augmentation techniques including synonym replacement, random deletion, word swapping, and named entity substitution. Our augmented S-BERT model achieved an accuracy of 85.0%, outperforming the non-augmented baseline at 82.3%. This approach demonstrates the effectiveness of language-driven representations in speech-based PD detection and highlights the value of augmentation in enhancing transformer-based medical diagnostics.
Keywords
Acknowledgement
This work was supported by the Institute of Information & Communications Technology Planning & Evaluation(IITP)-Innovative Human Resource Development for Local Intellectualization program grant funded by the Korea government(MSIT)(IITP-2025-RS-2022-00156287, 33%). This work was supported by Institute of Information & communications Technology Planning & Evaluation (IITP) under the Artificial Intelligence Convergence Innovation Human Resources Development (IITP-2023-RS-2023-00256629, 33%) grant funded by the Korea government(MSIT). This research was supported by the MSIT(Ministry of Science and ICT), Korea, under the ITRC(Information Technology Research Center) support program(IITP-2024-RS-2024-00437718, 34%) supervised by the IITP(Institute for Information & Communications Technology Planning & Evaluation).