Szczegóły publikacji
Opis bibliograficzny
An explainable AI-integrated diagnostic system for voice analysis in heart failure patients / Mikołaj Najda, Miłosz DUDEK, Olgierd Unold, Tomasz Jadczyk, Krzysztof Świerz, Grzegorz Świątek, Daria HEMMERLING // W: AAAI Bridge Program on AI for Medicine and Healthcare [Dokument elektroniczny] : 25 February 2025, Philadelphia, Pennsylvania, USA : proceedings / eds. Junde Wu, [et al.]. — Wersja do Windows. — Dane tekstowe. — [Cambridge : JMLR], cop. 2025. — ( Proceedings of Machine Learning Research ; ISSN 2640-3498 ; vol. 281 ). — S. [56]–[62]. — Wymagania systemowe: Adobe Reader. — Bibliogr. s. [61]–[62], Abstr. — T. Jadczyk – afiliacja: Medical University of Silesia, Katowice, Poland ; St. Anne’s University Hospital in Brno, Czech Republic
Autorzy (7)
- Najda Mikołaj
- AGHDudek Miłosz
- Unold Olgierd
- Jadczyk Tomasz
- Świerz Krzysztof
- Świątek Grzegorz
- AGHHemmerling Daria
Dane bibliometryczne
| ID BaDAP | 159900 |
|---|---|
| Data dodania do BaDAP | 2025-06-10 |
| Tekst źródłowy | URL |
| Rok publikacji | 2025 |
| Typ publikacji | materiały konferencyjne (aut.) |
| Otwarty dostęp |  |
| Czasopismo/seria | Proceedings of Machine Learning Research |
Abstract
Integrating Explainable Artificial Intelligence to analyse voice characteristics is an essential topic for future research. We explore the utility of tree-based machine learning models, including Random Forest, XGBoost, and LightGBM, in dis- tinguishing between two groups: 100 participants with heart failure and 100 healthy controls. The acoustic features ex- tracted from sustained vowel recordings are used to differentiate between the two groups. The evaluation shows that the Random Forest model performs better, especially with the vowel, achieving Accuracy, Precision, Recall, and F1 score over 0.80. We investigate the interpretability of these models using SHapley Additive exPlanations values, which reveal the essential acoustic features that influence model predictions and provide insights into their clinical relevance. This research highlights the potential of interpretable vocal biomarkers in remote monitoring and diagnosing heart failure.
