| 摘要 |
PROBLEM TO BE SOLVED: To provide a proper name associating method capable of widely acquiring the proper name equivalency by various methods, and acquiring a occurrence position of the proper name in an example for translation. SOLUTION: In this proper name associating method for acquiring the proper name equivalent between two languages from the example for translation as a pair of a first natural language sentence and a second natural language sentence as the translation of the first natural language sentence, the proper name in the first natural language sentence is extracted from the analysis information of the first natural language sentence even when the analysis information of the second natural language sentence does not exist, further the dictionary-equivalent, character-equivalent, reading- equivalent, alphabet-equivalent, dictionary-character equivalent data between the two languages are properly combined to estimate the words and phrases equivalency between the two languages, and the data on the proper name equivalency between the two languages and the occurrence position in each sentence are extracted from the proper name in the first natural language sentence and the words and phrases equivalency between two languages. |
