音译涉及的两种语言采用不同的字母表和发音系统时(英语和汉语,英语和日语,英语和阿拉伯语等),机器音译就更复杂且更具有挑战性。音译单元对齐结果的好坏严重影响机器音译的准确率,为此研究了英汉机器音译中音译单元的对齐,提出了基于音节首字母匹配的音译单元对齐方法,该方法在音译单元的对齐中有较好的表现。
It is very complicated and challenging to translate names and technical terms across languages that employ very different alphabets and sound systems, such as English/Chinese, English/Japanese and Arabic/English and so on. The quality of transliteration units alignment seriously affects the precision of Machine Transliteration. In our research, we concentrate on the transliteration unit alignment and propose a novel method for transliteration unit alignment which is called transliteration unit alignment based on the first letter. This method has a good performance in transliteration alignment.