INFORMATION COLLECTION SYSTEM, NATURAL LANGUAGE PROCESSING SYSTEM, AND MACHINE TRANSLATION SYSTEM

摘要

PURPOSE:To extract, and also, utilize the information different from conventional information, which can be utilized for a natural language processing and a machine translation by dispensing with other work than an input work and without a restriction related to an input sentence. CONSTITUTION:Such a conception as a long unit word consisting of a character string of two morphemes or more placed between a function word and a delimiter symbol is introduced. From an input sentence, information of the long unit word, bilingual translation information of the long unit word, an attribute of the word for constituting the long unit word or a translate word of the word constituting the long unit word, etc., is extracted and stored automatically 12-14 and 16. The extracted information is utilized suitably by a natural language processing 15.