| 摘要 |
PROBLEM TO BE SOLVED: To automatically acquire a group of translation rules from a document database. SOLUTION: This translation rule extracting program is prepared for operating a computer as a device for extracting a group of translation rules from a parallel corpus, and this device comprises a synonymous sentence collection forming part 110 for forming the synonymous sentence collection from the parallel corpus, a synonymous expression pair extracting part 112 for extracting the synonymous expression pair from each of the synonymous sentence collections, and a classifying part 116 for classifying the extracted synonymous expression pair into synonymous expression clusters, and creating a group of translation rules wherein each synonymous expression cluster is applied as the translation rule. The device may include a representative expression deciding part 118 for deciding the representative synonymous expression to each of the group of translation rules, and a processing part 114 for filtering the extracted synonymous expression pair on the basis of the frequency of appearance or the cooccurrence relation. COPYRIGHT: (C)2004,JPO
|
