| Bilingual Named Entity pairs are valuable resources for many NLP applications. Since comparable corpora are more accessible, abundant and up-to-date, recent researches have concentrated on mining bilingual lexicons using comparable corpora. Leveraging comparable corpora, this research presents a novel approach to mining English-Chinese Named Entity translations by combining multi-dimension features from various information sources for every possible Named Entity pair, which include Transliteration, English-Chinese matching, Chinese-English matching, Translation model, Length and Context Information. These features are integrated into one model with linear combination and minimum'sample risk (MSR) algorithm. As for the high type-dependence of Named Entity translation, we integrate different features according to different Named Entity types. We experiment with the above individual feature or integrated features to mine person Named Entity pairs, location Named Entity pairs and organization Named Entity pairs. When using transliteration and length to mine person Named Entity pairs, it achieves the best performance of 84.9%(F-score). The location Named Entity pairs can be mined with the features of transliteration model, length, translation model, English-Chinese matching and Chinese-English matching. And the best performance is 83.4% (F-score). The organization Named entity pairs can be mined with the features of English-Chinese matching and Chinese-English matching. It reaches the best performance of 84.1%(F-score). |
