Skip to main content
. 2021 Jan 25;2020:1041–1049.

Figure 2:

Figure 2:

Proposed Model: Each electronic health record (EHR) is split up into multiple paragraphs which are passed through the BERT Model. BERT predicts the relevant paragraphs which are used to extract textual features such as n-grams and TF-IDF. These features are then passed through a statistical model such as logistic regression to predict the causal relation label.