Skip to main content
. 2021 Oct 30;10:285. doi: 10.1186/s13643-021-01763-w

Fig. 1.

Fig. 1

Procedure of building the srBERT model using datasets obtained via previous SRs. The abstracts of documents downloaded in Endnote are used to create the model vocabulary and pre-train the model. Data categorized as “Title,” which were obtained through manual screening, were used for the fine-tuning of srBERT. SR, systematic review; BERT, Bidirectional Encoder Representations from Transformers