Skip to main content

Table 3 Comparison of performance metrics for the HEA BioBERT architecture at 80% and 100% coverage. Coverage refers to the percent of articles the model makes a prediction for (as opposed to abstaining from making a prediction when the model has a confidence below the Threshold). The Precision, Recall, and Accuracy scores reflect the accuracy of the model on the resulting 80% of predicted articles

From: AutoDiscern: rating the quality of online health information with hierarchical encoder attention-based neural networks

QuestionCoverageThresholdPrecisionRecallAccuracy
Q4: References80%0.790.870.7987%
Q5: Date80%0.790.870.8887%
Q9: How Treatment Works80%0.810.840.7182%
Q10: Treatment Benefits80%0.700.660.5583%
Q11: Treatment Risks80%0.860.900.9091%
Q4: References100%0.500.830.8084%
Q5: Date100%0.500.830.8383%
Q9: How Treatment Works100%0.500.770.7278%
Q10: Treatment Benefits100%0.500.570.5477%
Q11: Treatment Risks100%0.500.810.8181%