Comparative analysis of machine learning algorithms for computer-assisted reporting based on fully automated cross-lingual RadLex mappings
| Autor: | Chang Gyu Cho, Victor Saase, Fabian Siegel, Máté E. Maros, Frederik Trinkmann, Andreas G. Junge, Thomas Ganslandt, Benedikt Kämpgen, Holger Wenz, Christoph Groden |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2021 |
| Předmět: |
Elastic net regularization
Statistical methods Calibration (statistics) Computer science Science Linear classifier Machine learning computer.software_genre Predictive markers Article 030218 nuclear medicine & medical imaging 03 medical and health sciences Prognostic markers 0302 clinical medicine Classifier (linguistics) Computational models Data mining Multidisciplinary business.industry Probabilistic logic Diagnostic markers Computational biology and bioinformatics Support vector machine Tree (data structure) Brier score 030220 oncology & carcinogenesis Medicine Artificial intelligence business Algorithm computer |
| Zdroj: | Scientific Reports Scientific Reports, Vol 11, Iss 1, Pp 1-18 (2021) |
| ISSN: | 2045-2322 |
| Popis: | Computer-assisted reporting (CAR) tools were suggested to improve radiology report quality by context-sensitively recommending key imaging biomarkers. However, studies evaluating machine learning (ML) algorithms on cross-lingual ontological (RadLex) mappings for developing embedded CAR algorithms are lacking. Therefore, we compared ML algorithms developed on human expert-annotated features against those developed on fully automated cross-lingual (German to English) RadLex mappings using 206 CT reports of suspected stroke. Target label was whether the Alberta Stroke Programme Early CT Score (ASPECTS) should have been provided (yes/no:154/52). We focused on probabilistic outputs of ML-algorithms including tree-based methods, elastic net, support vector machines (SVMs) and fastText (linear classifier), which were evaluated in the same 5 × fivefold nested cross-validation framework. This allowed for model stacking and classifier rankings. Performance was evaluated using calibration metrics (AUC, brier score, log loss) and -plots. Contextual ML-based assistance recommending ASPECTS was feasible. SVMs showed the highest accuracies both on human-extracted- (87%) and RadLex features (findings:82.5%; impressions:85.4%). FastText achieved the highest accuracy (89.3%) and AUC (92%) on impressions. Boosted trees fitted on findings had the best calibration profile. Our approach provides guidance for choosing ML classifiers for CAR tools in fully automated and language-agnostic fashion using bag-of-RadLex terms on limited expert-labelled training data. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink