摘要 |
PROBLEM TO BE SOLVED: To more accurately and more speedily pre-edit a natural language sentence to be applied to a natural language processing system. SOLUTION: This device is composed of a morpheme information table storing morpheme information containing a vocabulary attribute such as part-of- speech information concerning the word of a natural language, pre-editing rule table storing the pre-editing rules of a natural language sentence based on the morpheme information, input part for inputting the natural language and a pre-editing instruction thereof, morpheme analytic part for performing morpheme analysis to the respective words of the inputted natural language sentence and extracting the morpheme information of each of words, pre-editing part for judging whether the inputted natural language sentence is the sentence of a special expression or not based on the morpheme information provided from the morpheme analytic result and the pre-editing rules in the pre-editing rule table, reloading the sentence of the special expression to the sentence of a standard expression and pre-editing that sentence, and output part for outputting the pre-edited natural language sentence. |