Table 1 The performances and comparisons between DL and rule based machine translation of chemical nomenclature
From: Neural machine translation of chemical nomenclature between English and Chinese
Field | CNN based | LSTM based | Rule based |
|---|---|---|---|
Success Rate En2Ch | 100% | 100% | 75.97% |
Success Rate Ch2En | 100% | 100% | 59.90% |
String Matching Accuracy En2Ch | 82.92% | 89.64% | 39.81% |
String Matching Accuracy Ch2En | 78.11% | 55.44% | 43.77% |
Data Matching Accuracy En2Ch | 84.44% | 90.82% | 45.15% |
Data Matching Accuracy Ch2En | 80.22% | 57.40% | 44.91% |
Manual Spot Check En2Ch | 90.00% | 89.00% | 80.00% |
Manual Spot Check Ch2En | 82.00% | 61.00% | 78.00% |
Running Time En2Ch (s) | 1423 | 190 | 288 |
Running Time Ch2En (s) | 1876 | 303 | 322 |