发明名称 METHOD AND DEVICE FOR ALIGNING TWO-LANGUAGE CORPUS
摘要 <p><P>PROBLEM TO BE SOLVED: To provide a method for aligning a sentence of a first corpus into a sentence of a second corpus. <P>SOLUTION: This method includes formation of an aligned pair of sentences by aligning a sentence boundary of sentences of a first corpus into a sentence boundary of sentences of a second corpus by adapting an alignment model based on length. Next, using the aligned pair of sentences, a translation model is trained. After finishing the training, using the translation model, a sentence of the first corpus is aligned into a sentence of the second corpus. According to a mode of this invention, using pruning, the number of sentence boundary alignments considered by the alignment model based on the length and the number of sentence boundaries considered by the translation model are reduced. In a further mode of the invention, the model based on the length uses Poisson distribution. <P>COPYRIGHT: (C)2004,JPO</p>
申请公布号 JP2004086913(A) 申请公布日期 2004.03.18
申请号 JP20030302014 申请日期 2003.08.26
申请人 MICROSOFT CORP 发明人 MOORE ROBERT C
分类号 G06F17/27;G06F17/28;(IPC1-7):G06F17/28 主分类号 G06F17/27
代理机构 代理人
主权项
地址