JPH0630098B2 - Kana-Kanji converter - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a kana-kanji conversion device for converting a Japanese sentence input by a kana code into a kana-kanji mixed sentence and outputting the sentence.

[Prior art]

A conventional kana-kanji conversion device (for example, Japanese Patent Laid-Open No. 57-1272).
No. 67), if the Japanese dictionary is searched and there is no corresponding word, or if there is a grammatical error due to an input error or the like, all kana-kanji conversion processing will result in an error, or at least the applicable phrase and word level. There was an error. However, when performing multiple bunsetsu conversion, etc., a long kana character string was input, and it was very uneconomical to make an error at least at the bunsetsu and word levels due to some input errors.

[Purpose] In view of such a problem, the present invention is an economical method that analyzes the remaining character string as effectively as possible with a minimum correspondence when it is not in the corresponding word dictionary, and minimizes the re-input operation. The purpose is to provide a kana-kanji conversion device that enables input.

[Characteristics of the Invention]

The present invention is a kana-kanji conversion device that analyzes kana-character strings input by key input or the like and performs kana-kanji conversion, and a phrase that generates a phrase candidate by searching a dictionary based on the input kana-character string. A candidate generation means, a clause end word test means for grammatically analyzing a connection between the generated clause candidate and a word located at a delimiter position of a clause to which the clause candidate belongs, which is a rear part of the clause candidate; An indefinite word generating means for extracting a word determined to be not grammatically connected to the bunsetsu candidate as an indefinite word by the bunsetsu ending word testing means and continuing analysis of the remaining input kana character string. Characterize.

〔Example〕

Here, the details of the present invention will be described below based on illustrated embodiments.

FIG. 1 is a schematic block diagram showing an embodiment of the kana-kanji conversion device of the present invention. Reference numeral 1 in the figure is a data input unit such as a keyboard. Of the data input by the data input unit 1, only the Kana data is accumulated in the Kana character sentence receiving / storing unit 2, and by the control information of the analysis start by the data input unit 1, the bunsetsu candidate expanding means 3 causes the kana digit to expand. Position 1
Starting from the digit, the national language dictionary searching means 4 indexes the national language dictionary 5 from that position, and the bunsetsu candidate developing means 3 stores the bunsetsu candidates in the bunsetsu candidate storage unit 7. At this time, the bunsetsu candidates are candidates. The ending digit position of is also stored. Then, 1 is added to the kana digit position of the base point, and the next base point analysis is started. At this time, if there is no digit position at the end of the phrase after this base point,
The indefinite word generating means 6 generates a one-digit kana indefinite phrase and sets this digit as the end of the phrase. At this time, when the bunsetsu candidates are expanded and stored in the bunsetsu candidate buffer 8, the bunsetsu candidate joining verification means 9 checks the grammatical information by the grammar checking unit 10, and the grammatical error data is stored in the bunsetsu candidate buffer 8. Set an error flag on the data. Further, when this processing is completed, the document candidate re-editing means 11 verifies the data in the phrase candidate buffer 8 whether or not the entire combined state can be maintained without an indefinite word, and if that is possible, the indefinite word data. Is deleted as an error together with the data in error by the document candidate joining verification means 9, and valid data is reorganized in the phrase candidate buffer 8.

When these series of operations are completed, the phrase candidate output means 1
2 uses the data of the reorganized phrase candidate buffer 8 as C
Output to RT etc.

Next, the operation of the embodiment of FIG. 1 will be described in more detail with reference to the flowchart of FIG.

First, the value of a kana character digit position pointer (not shown) indicating the digit position of the kana character string input by the data input unit 1 and accumulated in the kana character sentence reception storage unit 2 is set to 1 (<1 >). Next, the kana character digit position pointer and the number of digits of the input kana character string are compared (<2>), and if the value of the kana character digit position pointer is less than or equal to the value of the input kana character string, the processing of <3> is performed. If the value of the kana character string pointer becomes larger than the value of the input kana character string, the processing described in <10> and below will be performed.

In <3>, the national language dictionary searching means 4 indexes the national language dictionary 5, and the bunsetsu candidate expanding means 3 further sets the bunsetsu candidate storing section 7.
The bunsetsu candidate is stored in, and the ending digit position of the bunsetsu candidate is also stored. This point will be described in more detail with reference to FIG.

In FIG. 3, 30 is a kana character string input from the data input unit 1, 31 is a phrase end digit table corresponding to the digit position, and is a place corresponding to the digit position where the word at the end of the sentence exists. It stands out.

Now, giving an example, first, I found the independent word "I" (to be precise, there is an independent word "wa", but for the sake of explanation, I will explain it myself). "Ha" is found as the ending that follows,
The word "I am" is cut out. Then, it is verified by the bunsetsu candidate expanding means 3 whether this bunsetsu end is correct, and if the correct word continues as the bunsetsu end as in the present case, a flag is set in the bunsetsu end digit table at the corresponding digit position. However, for example, if "n" is input instead of "ha", the word at the end of the phrase is determined to be incorrect by the test at the end of the phrase, and no flag is set in the end-of-segment digit table. .

Next, returning to the flowchart of FIG. 2 again, the description of the operation is continued. After <3>, it is checked whether the phrase end table flag is set at the nearest backward phrase end position (the position of “ha” in the example of FIG. 3) closest to the stem of the cut out independent word or the like. (<4>), if the flag is set,
The value obtained by adding one to the digit position where the flag is set is set as the value of the kana character digit position pointer (<5>), and the process returns to <2>. If the flag is not set in the process of <4>, the indefinite word generating means 6 generates an indefinite word of one character (<6>). That is, the kana character at the digit position where the flag should have been set is regarded as an indefinite word. After <6>, the value of the kana character digit position pointer becomes a value obtained by adding +1 to the value up to that point (indicating the digit position of the indefinite word) (<7>), and then the processing of <2>. Return to.

When the value of the kana character digit position pointer exceeds the number of input digits of the kana character string in the processing of <2>, it is next determined whether the word at the end of the sentence is correct as the word at the end of the sentence (<10>), if it is not correct, the word at the end of the sentence is set as an indefinite word (<11>), and the process is ended.

FIG. 4 is an example in which an indefinite word is generated as a candidate by the indefinite word generating means 6 by the document candidate expanding means 3 in FIG. For example, in the case of making a relatively mistaken input, "Kyouhatenkiga" should be entered as "Kyouhatenkiga"
The case where it is input is given.

However, for the sake of convenience, it is assumed that only those that have been filled in are registered in the national language dictionary used at this time.

In this case, the second and third characters "yo" and "u" occur as indefinite words in units of one character.

According to this, it can be seen that the analysis of “Kyouha” corresponding to “Kyoha” in the first half has failed, but the analysis of the part corresponding to “Tenkiga” in the latter half has succeeded. .

〔effect〕

As described above, in the present invention, the connection between the generated phrase candidate and the word located at the delimiter position of the phrase to which the phrase candidate belongs is grammatically analyzed,
Since the indefinite term is extracted based on the result, the indefinite term can be extracted only by examining only a few combinations of words and phrases, and the indefinite term can be extracted easily and in a short time.

Also, because it is based on grammatical analysis,
Indefinite words can be extracted even when a list of words that are not grammatically connected (for example, a case where individual words are in a dictionary such as “I”) is extracted.

[Brief description of drawings]

FIG. 1 is a functional block diagram showing an outline of one embodiment of the kana-kanji conversion device of the present invention. FIG. 2 is a block flowchart of the operation of the embodiment shown in FIG. FIG. 3 is a diagram for explaining a part of the block flowchart of FIG. 2 in more detail. FIG. 4 is a specific example of expanding the clause candidates to which the present invention is applied. 3 ... clause candidate expansion means 6 ... indefinite word generation means

Claims (1)

[Claims] 1. A kana-kanji conversion device that analyzes kana-kanji character strings input by key input or the like and performs kana-kanji conversion, searches a dictionary based on the input kana-character strings to generate phrase candidates. A clause candidate generating means, a clause ending word test means for grammatically analyzing a connection between the generated clause candidate and a word located at a delimiter position of a clause to which the clause candidate belongs, which is a rear part of the clause candidate. An indefinite word generating means for extracting a word determined to be not grammatically connected to the bunsetsu candidate as an indefinite word by the bunsetsu end word testing means and continuing the analysis of the remaining input kana character string. Kana-to-Kanji conversion device featuring.