Table 4.
Topic classification for n = 1,074 annotations. Classifiers are listed by their highest F-measure. IAA indicates the rank order of the topic IAA scores. Bold indicates highest performance between approaches.
| Topic | IAA | Precision | Recall | F-measure | |||
|---|---|---|---|---|---|---|---|
| Vocabulary | Naive Bayes | Vocabulary | Naive Bayes | Vocabulary | Naive Bayes | ||
| Explanation | 1 | 0.42 | 0.90 | 0.35 | 0.63 | 0.38 | 0.74 |
| Appointment wait | 4 | 0.31 | 0.54 | 0.32 | 0.60 | 0.31 | 0.57 |
| Appointment access | 2 | 0.45 | 0.44 | 0.46 | 0.64 | 0.46 | 0.52 |
| Practice environment | 7 | 0.38 | 0.24 | 0.61 | 0.57 | 0.48 | 0.34 |
| Overall | 3 | 0.66 | 0.85 | 0.30 | 0.27 | 0.44 | 0.41 |
| Friendliness | 5 | 0.50 | 0.32 | 0.31 | 0.57 | 0.39 | 0.40 |
| Empathy | 6 | 0.41 | 0.18 | 0.33 | 0.14 | 0.36 | 0.16 |