| 摘要 |
PROBLEM TO BE SOLVED: To provide a method and a device for analyzing a Japanese text enabling to conduct correct read-aloud of the text even when the inputted text contains a Roman character notation text unregistered in a word dictionary. SOLUTION: When the inputted text contains a text of Roman character notation (step 200, 201), the word dictionary is retrieved (step 202, 203). When the corresponding word is found, the text is replaced with a phonogram string by referring to the word dictionary. When the Roman character notation text is not registered in the word dictionary, if the text is the Roman character notation of a Japanese word (step 204, 205), it is outputted as a phonogram string of syllable unit (step 205a) and if not, the text is converted into a phonogram string obtained by reading the spell of each Roman character (step 206). |
