Summary of evaluation results of a heuristic model, three conventional machine learning models (LR, SVM, and GB), a neural networks model (CNN), and three BERT-based models (BERT, BioBERT, and SciBERT) for using (a) title, (b) abstract, and (c) combined title and abstract as input data.

Bold indicates the best performance results among all models. LR: Logistic Regression, SVM: Support Vector Machines, GB: Gradient Boosting, CNN: Convolutional Neural Networks.