A Report On The Comparison Of The Efficiency Of Automatic Bilingual Term Extraction Tools

Efficient computer-aided translation depends heavily on high-quality translation memories and termbases.Automatic extraction of bilingual terms is a key step in building a large-scale bilingual termbase.To explore the automatic extraction method of bilingual terms and improve the existing automatic extraction tools of bilingual terms have become important contents of computer-aided translation research.In this study,"Maritime Safety English-Chinese Parallel Corpus," is used as the experimental corpus to compare the accuracy,recall rate and F value of the extraction results of SDL Multi Term Extract 2019,Sketch Engine One Click Terms and Synchro Term by the method of bilingual term extraction experiment.On this basis,the influence of the translation direction,length and frequency of the extracted terms on the extraction efficiency of each tool is further discussed.Experimental results show that the overall efficiency of the three tools for extracting English-Chinese bilingual terms is higher than that of Chinese-English bilingual terms.In terms of specific tools,Synchro Term has the highest overall extraction efficiency,followed by SDL Multi Term Extract 2019,and Sketch Engine One Click Terms the lowest.And for extracting bilingual terms of different length and frequency,Synchro Term has the best overall extraction effect,while SDL Multi Term Extract 2019 and Sketch Engine One Click Terms only have advantages for term extraction in specific lengths and frequency bands.The results of this experiment will provide a reference for translators or translation teams and point out the direction for the optimization of bilingual term extraction tools.