| 摘要 |
PURPOSE: To reduce the diversity of division and that of meaning to perform accurate analysis by detecting the error of readings of FURIGANA (phonetic transcriptions in Japanese syllabary) assigned to individual words and substituting the reading with a correct reading or expression at the time of detecting the error of the reading to settle segmentation and the reading of a segmented word. CONSTITUTION: An inputted expression is divided into words by an expression, analysis part 304, and the analysis result is stored as a solution candidate in a temporary storage part 305. If there are plural analysis results, they are stored in plural temporary storage parts 305. Next, a FURIGANA analysis part 308 takes out candidates as analysis results from temporary storage parts 305 one by one to analyze FURIGANA. A character dictionary 307 is retrieved by the first word to check whether the dictionary has its characters and reading or not. If they are found, auxiliary information is referred to perform diversity solution of division, that of meaning, correction of the expression, etc. If they are not found, another candidate is analyzed. When all of the character string is analyzed, this analysis is terminated. |
