JPH0394367A - Japanese input system - Google Patents

Abstract

PURPOSE:To prevent a learning dictionary from being reduced due to its overflow by automatically registering the learning dictionary in a user dictionary by frequency control. CONSTITUTION:A 'KANA' (Japanese syllabary)/'KANJI'(Chinese character) conversion dictionary consists of a work dictionary 11 and a learning area 12, the contents of the word dictionary 11 are constituted of a 'KANA' header and a 'KANJI' notation as shown in the reference code 13 and the contents of the learning area 12 are constituted of a 'KANA' header, a 'KANJI' notation and using frequency as shown in the reference code 14. The using frequency of words in the learning area 12 is counted and stored and the words with high using frequency are registered in the word dictionary. When a register are overflows, the area can be secured by deleting words with lower using frequency. Consequently, the words in the learning dictionary can be prevented from being automatically deleted due to the overflow of the learning area.

Description

[Detailed description of the invention] [Industrial application field] The present invention relates to a Japanese input method.

[Conventional technology]

Traditionally, in Kana-Kanji conversion using Japanese input methods, character strings input in Kana are converted into kanji-kana mixed sentences using a word dictionary consisting of words prepared in advance by the Japanese input system and words added by the user. I was doing this. [Problems to be Solved by the Invention] In the conventional Japanese input method that performs kana-kanji conversion as described above, input for words that are not registered in the word dictionary results in incorrect conversion, and the input method is divided into single kanji characters by a method such as re-separation. This is supplemented by using a method of converting to the corresponding word. With conventional Japanese input methods, the word dictionary is not updated when conversion is performed by re-dividing, etc., and frequently used words must be registered in the word dictionary each time, which hinders operability. It has the disadvantage of being a cause. [Means for Solving the Problems] The Japanese input method of the present invention inputs a character string expressed in kana, searches word dictionaries and learning areas as a heading, and uses the corresponding kanji notation to generate sentences containing kanji and kana. In the Japanese input method that performs kana-kanji conversion, if a word that matches the heading does not exist in the word dictionary, the heading is divided and searched in the word dictionary, and the kanji output for the divided heading is combined and the word a learning area for temporarily storing the headings before division and words after combination; a means for counting and storing the frequency of use of words in the learning area; and registering frequently used words in a word dictionary. and means.

〔Example〕

Next, the present invention will be explained with reference to the drawings.

FIG. 1 is an explanatory diagram of the kana-kanji conversion dictionary used in this embodiment. It consists of a word dictionary 11 and a learning area 12. The contents of the word dictionary 11 are composed of kana headings and kanji notations, as shown by reference numeral 13. The content of the learning area 12 is composed of a kana heading, a kanji notation, and a usage frequency, as shown by reference numeral 14.

FIG. 2 is an explanatory diagram of the kana-kanji conversion used in this embodiment. When the conversion 22 is instructed for the input kana heading 21, the kana-kanji conversion is performed using the word dictionary group, resulting in 23 converted kanji. Displays candidates. A confirmation instruction is given at step 24 for the correctly converted "Today's", and an instruction is given at step 26 to convert the erroneously converted "zeicho" shown at 25 into a kanji character. ``Zeichōj'' is converted into a single kanji character as ``zei'' (tax, force...) and ``cho''.
(major, key, intestinal, etc.) are displayed as candidates (27 and 29), and the operator selects ``tax'' and ``choice'' in 28 and 30 (31 and 32) and confirms them (in the operating steps up to 33). be.

3 to 5 are control flow charts.

Figure 3 shows the main control for kana-kanji conversion, which includes inputting headings (step 4) and performing kanji conversion (
Step 42>, grammar processing (step 43) for associating kanji with adjunct words, and confirmation processing (step 44) for displaying candidates and checking by the operator.

Figure 4 shows the main control of kanji conversion, in which the input heading is searched in the learning dictionary (step 51) and then searched in the word dictionary (step 53). 52.54), output the kanji notation (step 56).If there is a mismatch, output the hiragana as is (step 55).

FIG. 5 is a flowchart of the confirmation process, in which the output candidate display (step 61) is processed by the operator's key operation (step 61). The next candidate is an operation that requests the display of a homophone (step 62), and the next candidate word of the homophone is obtained by kanji conversion processing for this key and the candidate is displayed again (step 63), (step 64) The re-break is a key that instructs to shorten the heading and convert it to kanji.
The heading is shortened, the corresponding kanji is obtained from the kanji conversion process, and candidate kanji are displayed (step 65). If a key other than the above is pressed, this is an instruction to confirm the candidate display data and proceed to the next process, and the sentence containing kanji and kana is output as normal data (step 66).

At this time, if a kanji is selected by the re-separation instruction (step 67), the kana heading and the kanji are registered in the learning area (step 68).If the registration area overflows, The area is secured by deleting words that are used less frequently.If the words output from the learning area are used (step 69)
Increase the frequency information of the relevant heading by 1 (step 70),
Further, if the frequency information exceeds the system default value i (step 71), the corresponding heading is registered in the user dictionary so as not to be automatically deleted due to subsequent learning area overflow, and is deleted from the learning dictionary (step 72).

〔Effect of the invention〕

As explained above, the present invention converts incorrectly converted headings into words by dividing them and converting them into kanji, and then temporarily registers the headings and words in the learning area.
When the same heading is input for the second time or later, the corresponding word can be directly output, and the means for automatically registering it in the user dictionary through frequency control has the effect of preventing reductions in the learning dictionary due to overflow.

[Brief explanation of drawings]

Fig. 1 is an explanatory diagram of a kana-kanji conversion dictionary used in one embodiment of the present invention, Fig. 2 is an explanatory diagram of kana-kanji conversion used in this embodiment, and Fig. 3 is a flowchart of control of kana-kanji conversion of the embodiment. FIG. 4 is a flowchart of kanji conversion in this embodiment, and FIG. 5 is a flowchart of confirmation processing in this embodiment. 11...Word dictionary, 12...Learning area.

Claims (1)

[Claims] In the Japanese input method, which performs kana-kanji conversion, which inputs a character string expressed in kana, searches the word dictionary and study area as a heading, and converts it into a sentence containing kanji and kana using the corresponding kanji notation, the word that matches the heading If the heading does not exist in the word dictionary, there is a means to divide the heading and search the word dictionary, a means to combine the kanji output for the divided heading into a word, and a means to temporarily combine the heading before dividing and the word after combining. A Japanese language input method comprising a learning area to be stored, means for counting and storing the frequency of use of words in the learning area, and means for registering frequently used words in a word dictionary.