Repozytorium publikacji - Politechnika Gdańska

Proper triage of COVID-19 patients is a key factor in eective case management, especially with limited and insucient resources. In this paper, we propose a machine-aided diagnostic system to predict how badly a patient with COVID-19 will develop disease. The prognosis of this type is based on the parameters of commonly used complete blood count tests, which makes it possible to obtain data from a wide range of patients.We chose the four-tier nursing care category as the outcome variable. In this paper, we compare traditional tree-based machine learning models with approaches based on neural networks. The developed tool achieves a weighted average F1 score of 73% for a three-class COVID-19 severity forecast. We show that the complete blood count test can form the basis of a convenient and easily accessible method of predicting COVID-19 severity. Of course, such a model requires meticulous validation before it is proposed for inclusion in real medical procedures.

Autorzy

• Barbara Klaudel link otwiera się w nowej karcie ,
• Aleksander Obuchowski link otwiera się w nowej karcie ,
• Roman Karski link otwiera się w nowej karcie ,
• Bartosz Rydziński link otwiera się w nowej karcie ,
• dr inż. Patryk Jasik link otwiera się w nowej karcie ,
• prof. dr hab. inż. Zdzisław Kowalczuk link otwiera się w nowej karcie