| 摘要 |
<p><P>PROBLEM TO BE SOLVED: To provide a method and apparatus for training a parsing statistical model and a method and apparatus for transliteration. <P>SOLUTION: The parsing statistical model is used for transliteration between a monosyllabic language and a polysyllabic language and includes the sub syllable parsing probability of the polysillabic language. In the method for training the parsing statistical model, a bilingual intrinsic personal name list is input as a corpus, and the bilingual intrinsic personal name list includes a plurality of intrinsic personal names of the polysillabic language and the corresponding intrinsic personal names of the monosyllabic language respectively. In the method, the plurality of intrinsic personal names of the polysillabic language inside the bilingual intrinsic personal name list are parsed to sub syllable columns respectively by using the rules of parsing, whether or not parsing is accurate is determined according to the corresponding intrinsic personal names of the monosyllabic language inside the bilingual intrinsic personal name list, and the parsing statistical model is trained on the basis of a parsed result determined to be accurate. <P>COPYRIGHT: (C)2007,JPO&INPIT</p> |
