| 摘要 |
Disclosed are a method and an apparatus for normalizing a natural language wherein natural language data are clustered in units for performing similar functions, and a normalization rule is generated by using a normalization word selected among normalization candidates extracted from a clustering result based on similarity. The method for normalizing the natural language comprises: a pre-processing step for generating the natural language data; a similarity generating step for generating a similarity list; a candidate processing step for extracting normalization candidates; a normalization control step for selecting a normalization word selected among normalization candidates; and a normalization rule generating step. |
