| 摘要 |
PURPOSE:To obtain information for efficiently and exactly customizing data for natural language processing. CONSTITUTION:A sample text extraction part 1 extracts a sample text which is proper in quantity or quality from a corpus 7 of a field to be processed. The extracted text may be stored in a sample file 5. A user indicates analytic information for the sample text through an indication part 2. Proper information is predetermined as the analytic information corresponding to the feature of a language to be processed and the ability of the user. Further, information in a general dictionary file 8 can be referred to when necessary. The indicated information is analyzed at an analysis part 3 to obtain the information required to customize a dictionary and grammar data. After a process for frequency calculation is performed at a storage part 4, the result is stored in an analysis result storage file 6. |
