Skip to main content
. 2022 Sep 14;13:946387. doi: 10.3389/fpsyt.2022.946387

TABLE 10.

Comparison of NLP performance for analysis of specific pathologies.

Author Analyzed pathology NLP model Embeddings and Corpus F1-score [%]
Castro et al. (107) Cerebral aneurysms cTAKES-based + LASSO Train corpus: ∼300 clinical notes (manually annotated)
Validation corpus: ∼17 k clinical notes
84
Katsuki et al. (108) Primary headache DL-based Test corpus: ∼850 clinical notes 63.5

NLP, natural language processing; cTAKES, clinical text analysis and knowledge extraction system; LASSO, L1-regularized logistic regression; DL, deep learning.