Study develops a new algorithm for assessing clinical trial abstracts.
problem Limited tools for assessing qualitative statements in clinical research.
method Developed a three-class sentiment classification algorithm using BERT model.
result Algorithm achieved 91.3% classification accuracy and 0.92 macro F1-Score.