TABLE VI.
Accuracy of cross-utterance classification and cross-validation within utterances showing the effect of increasing the amount of training data. Data were from exp. 3. Abbreviations: Talker (T), utterance (U), same (S), and different (D). In the case of the different-utterance condition, all utterances were still of the same word. Bold entries are significantly better than chance at a 0.05 level according to a one-sided binomial test.
| T | U | Ntr | /ɑtʃɑ/ | /ɑdʒɑ/ | /ɑdɑ/ | /ɑtɑ/ | /ɑfɑ/ | /ɑvɑ/ | Avg. |
|---|---|---|---|---|---|---|---|---|---|
| S | S | 104.9 | 59.0 | 66.5 | 60.9 | 59.3 | 58.7 | 56.8 | 60.2 |
| S | D | 105.1 | 57.6 | 62.0 | 69.5 | 63.7 | 56.5 | 60.0 | 61.5 |
| S | D | 262.2 | 65.2 | 62.7 | 74.1 | 74.4 | 67.0 | 63.5 | 67.8 |
| D | S | 114.3 | 57.9 | 59.2 | 60.7 | 61.9 | 54.1 | 55.4 | 58.2 |
| D | D | 114.1 | 50.7 | 49.6 | 54.8 | 59.0 | 55.1 | 53.9 | 53.8 |
| D | D | 464.7 | 62.3 | 58.7 | 69.3 | 73.9 | 59.9 | 61.6 | 64.3 |