Cargando…
Domain Shifts in Machine Learning Based Covid-19 Diagnosis From Blood Tests
Many previous studies claim to have developed machine learning models that diagnose COVID-19 from blood tests. However, we hypothesize that changes in the underlying distribution of the data, so called domain shifts, affect the predictive performance and reliability and are a reason for the failure...
Autores principales: | Roland, Theresa, Böck, Carl, Tschoellitsch, Thomas, Maletzky, Alexander, Hochreiter, Sepp, Meier, Jens, Klambauer, Günter |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
Springer US
2022
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8960704/ https://www.ncbi.nlm.nih.gov/pubmed/35348909 http://dx.doi.org/10.1007/s10916-022-01807-1 |
Ejemplares similares
-
Lifting Hospital Electronic Health Record Data Treasures: Challenges and Opportunities
por: Maletzky, Alexander, et al.
Publicado: (2022) -
Machine learning–based prediction of transfusion
por: Mitterecker, Andreas, et al.
Publicado: (2020) -
DEXUS: identifying differential expression in RNA-Seq studies with unknown conditions
por: Klambauer, Günter, et al.
Publicado: (2013) -
CLOOME: contrastive learning unlocks bioimaging databases for queries with chemical structures
por: Sanchez-Fernandez, Ana, et al.
Publicado: (2023) -
Large-scale comparison of machine learning methods for drug target prediction on ChEMBL
por: Mayr, Andreas, et al.
Publicado: (2018)