| 摘要 |
Disclosed are an apparatus for generating a self-learning alignment-based alignment corpus, a method therefor, an apparatus for analyzing a destructive expression morpheme by using an alignment corpus, and a morpheme analysis method therefor. The apparatus for analyzing the morpheme includes a knowledge database and an analyzer. The knowledge database stores a large amount of knowledge information used for analyzing morphemes of each language, and includes a morpheme dictionary for storing morpheme information corresponding to normal expressions and an alignment corpus for storing normal expression information corresponding to destructive expressions - here, the destructive expressions are expressions including spelling errors or expressions which are not normalized or standardized. The analyzer performs morpheme analysis for an inputted syntactic word by using the knowledge database and outputs the analyzed result, and when the morpheme for the inputted syntactic word does not exist in the morpheme dictionary, performs morpheme analysis by using the alignment corpus with respect to the destructive expressions included in the inputted syntactic word so as to find the normal expressions corresponding to the destructive expressions. |
