Szczegóły publikacji
Opis bibliograficzny
Machine learning-based decision support for allergy diagnosis: real-world implementation in a hospital setting / Paulina Tworek, Maja Szczypka, Julia Kahan, Marek Mikołajczyk, Roman Lewandowski, Jose Sousa // W: Artificial Intelligence in Medicine : 23rd international conference, AIME 2025 : Pavia, Italy, June 23–26, 2025 : proceedings , Pt. 1 / eds. Riccardo Bellazzi [et al.]. — Cham : Springer Nature Switzerland, cop. 2025. — ( Lecture Notes in Computer Science ; ISSN 0302-9743. Lecture Notes in Artificial Intelligence ; vol. 15734 ). — ISBN: 978-3-031-95837-3; e-ISBN: 978-3-031-95838-0 . — S. 448–456. — Bibliogr., Abstr. — Publikacja dostępna online od: 2025-06-23. — M. Szczypka, J. Kahan - dod. afiliacja: Sano – Centre for Computational Personalised Medicine - International Research Foundation, Krakow
Autorzy (6)
- Tworek Paulina
- AGHSzczypka Maja
- AGHKahan Julia
- Mikołajczyk Marek
- Lewandowski Roman
- Sousa Jose
Słowa kluczowe
Dane bibliometryczne
| ID BaDAP | 166229 |
|---|---|
| Data dodania do BaDAP | 2026-03-13 |
| DOI | 10.1007/978-3-031-95838-0_44 |
| Rok publikacji | 2025 |
| Typ publikacji | materiały konferencyjne (aut.) |
| Otwarty dostęp |  |
| Wydawca | Springer |
| Konferencja | Artificial Intelligence in Medicine 2025 |
| Czasopismo/seria | Lecture Notes in Computer Science |
Abstract
Up to 35% of children worldwide are affected by allergic diseases, and the incidence is steadily rising. Diagnosing these conditions is highly complex due to the wide range of allergens, the influence of cofactors, and symptoms that often do not unfold in medical tests. Moreover, allergies change with ageing, making rapid and accurate diagnoses fundamental to improving patient well-being while reducing the medical workload. Data-driven decision-making systems have long been presented as a highly effective solution. However, implementing decision-support solutions in hospital settings has several challenges, such as missing data, noise from varying lab procedures, and a lack of standardized diagnostic guidelines. To minimize the challenges and explore decision support potential, we conducted an experimental implementation of the Comprehensive Abstraction and Classification Tool for Uncovering Structures (CACTUS) in the Allergy Department at the Voivodeship Rehabilitation Hospital for Children in Ameryka, Poland. CACTUS was evaluated against explainable machine learning models, including Random Forest, Gradient Boosting, XGBoost, and AdaBoost. It outperformed the compared models by achieving superior diagnostic accuracy while consistently identifying the same key features influencing classification. The classification outcomes presented differences in clinicians’ diagnostic approaches, demonstrating CACTUS's potential to support clinicians’ individual decision-making processes. This real-world experiment shows the significant benefits of CACTUS as a decision-support tool in allergic diseases diagnostics by improving diagnostic accuracy, reducing unnecessary testing, and streamlining clinical workflows. CACTUS shows a strong potential to be a solution for allergic disease in-time diagnosis.
