Table 3 Comparison of models’ performance classification results. Model #1 was evaluated on the analyzed dataset. Model #2 was evaluated on the analyzed dataset, with the exception of two training examples that were used for few-shot learning. Model #3 was evaluated on the Test set of the analyzed dataset.
From: AI and narrative embeddings detect PTSD following childbirth via birth stories
Model | AUC | F1 score | Sensitivity (Recall) | Specificity |
|---|---|---|---|---|
Model #1 | 0.60 | 0.33 | 0.20 | 0.99 |
Model #2 | 0.60 | 0.38 | 0.24 | 0.96 |
Model #3 | 0.80 | 0.81 | 0.85 | 0.75 |
