| 摘要 |
PURPOSE:To increase the correct answer ratio of KANJI (Chinese character) conversion by string data on the KANJI mixed words together with the connection information between respective words. CONSTITUTION:In S1, the head of a character string to execute the conversion from a cursor address is retrieved, and in S2, a pointer is set to a head character position. In S3, a word which can be the last word of a segment, is retrieved with the longest coincidence method by referring to a dictionary memory. In S5, the pointer is updated, in S6, it is decided whether or not the word becomes the last of the segment. If the answer is NO, the next word is retrieved in the same way as previously, and in S8, the connection information with the previous word is decided on the basis of the connection information. When a word which can comes to be the last of the segment is found in S6, the word retrieving is executed concerning the next continuous character string in S9, and when in S11 and S12, it is decided that the word found by the word retrieving in S9 can be connected with the previous word and comes to be the head of the segment, the words, which are decided in S6 to form the segment are selected as the candidate character string in S13.
|
