Research Article
Unsupervised Quality Estimation Model for English to German Translation and Its Application in Extensive Supervised Evaluation
Table 4
The performances on WMT11 training corpora using Spearman correlation.
| System | Correlation score with human judgments | Other-to-English | English-to-other | Mean | CS-EN | DE-EN | ES-EN | FR-EN | Mean | EN-CS | EN-DE | EN-ES | EN-FR | Mean |
| | 0.95 | 0.61 | 0.96 | 0.88 | 0.85 | 0.68 | 0.35 | 0.89 | 0.83 | 0.69 | 0.77 | METEOR | 0.91 | 0.71 | 0.88 | 0.93 | 0.86 | 0.65 | 0.30 | 0.74 | 0.85 | 0.64 | 0.75 | BLEU | 0.88 | 0.48 | 0.90 | 0.85 | 0.78 | 0.65 | 0.44 | 0.87 | 0.86 | 0.71 | 0.74 | TER | 0.83 | 0.33 | 0.89 | 0.77 | 0.71 | 0.50 | 0.12 | 0.81 | 0.84 | 0.57 | 0.64 |
|
|
Bold fonts mean the best performance.
|