| 摘要 |
<p>PURPOSE:To improve the efficiency of conversion by segmenting candidate characters for conversion from the left end of an input character string, finding out its tolerance, executing processing while holding the candidate characters having tolerance included in a certain range, and after the end of the processing, outputting a candidate character string having the highest tolerance out of the candidate character strings for conversion and displaying also other candidate groups. CONSTITUTION:When a ''kana'' (Japanese syllabary) character string is inputted from an input part 1 and its conversion is specified, a control part 2 forms candidate character strings having high tolerance of probability of conversion and stores the candidate character strings in a storage part 3. Then the control part 2 forms and displays the candidate character string having the highest tolerance out of the candidate character strings stored in the storage part 3 and a substitutive candidate character string to be selected in accordance with the leading candidate character string, so that the user can select the objective candidate character. In case of clause conversion, a pointer indicating a character position counted from the leading of the ''kana'' character string is successively advanced and the number of clauses is found out on the basis of weight considering the parts of speech, appearance frequency, etc. It is regarded that the number of clauses is inversely proportional to the height of tolerance, and when the tolerance exceeds its allowable value, the clause is picked up and displayed.</p> |
