JPH0682367B2 - Document creation / proofreading support device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION <Industrial application field> The present invention extracts a part related to a proper noun from a Japanese sentence by performing a morphological analysis, and confirms and corrects the correctness of the proper noun. The present invention relates to a document creation / proofreading support device that facilitates performing.

<Conventional Technology> Currently, Japanese word processors have been put into practical use, and related basic technologies such as Japanese input / output, editing, Kana-Kanji conversion algorithm, and dictionary technology have been established.

In addition, basic Japanese processing techniques such as morphological analysis, syntactic analysis, and semantic analysis are known.

In Europe and the United States, word processors have developed from an early stage, so related technologies have advanced, and devices with spell check and collect functions have been put into practical use.

Compared to Western words that are written in units of words, Japanese is usually written in Kanji Kana altering sentences without punctuation. A device that is difficult to automate the calibration has not been put to practical use.

In the past, when it was required to handle accurate Japanese, multiple people would be paired and read to identify problematic points, or people with specialized knowledge of proofreading would be able to perform sequential matching. In addition, methods such as calibration method are adopted.

Recently, a device for supporting such calibration work is being developed. These devices can be roughly classified into two types.

One is for the former, which is a proofreading device capable of analyzing a sentence containing kanji and kana and converting it into a voice, and performing reading through voice with a machine.

The other is for the latter, with the purpose of supporting the automation of calibration or specialized calibration. Since this device requires higher technology than the former device, there are few reports of means for constructing the device and specific means of calibration only by the proposal of the concept.

In addition, as another technique, in language processing research, KW
The technology related to IC (Key Word In Context) is general.

Other than language processing, techniques related to workstations have been established, and devices using multi-windows have been put to practical use to improve work efficiency.

<Problems to be Solved by the Invention> In recent years, Japanese word processors have become widespread, and the number of documents created by the device has increased. In word processors, Kana-Kanji conversion method that is easy to input (including the Roman-Kanji conversion method unless otherwise specified) has been adopted in many models.

The kana-kanji conversion algorithm is based on a probabilistic basis, because it matches with linguistically-supported information such as word dictionaries and grammars in the process of converting kana to kanji. Has been confirmed.

Part of proper nouns is registered in a dictionary and often searched and confirmed in the same manner as a word dictionary.

However, since there are many proper nouns, it is not feasible to register all of them in the word dictionary in terms of economy, processing time, hardware restrictions, and the like.

Also, proper nouns often change, so be careful when updating them.

Also, if you make a mistake with a proper noun, it will have a great effect, and you will lose trust. Human rights and litigation issues may develop, and it is necessary to be careful in checking their correctness.

As described above, even in the case of a device having a proofreading function which is being developed recently, the function of checking and proofreading the proper noun part has not been reported.

For example, with a device that has a voice reading function, sentences are decomposed and converted into character strings one by one and read aloud by a voice synthesizer, so it is possible to perform strict confirmation even for proper noun parts, but processing speed Has the drawback of being slow.

In addition, since there are many proper nouns and letters with the same characters and different sounds, it was not always possible to proofread accurately even when reading with voice.

In addition, this device requires an additional device such as a voice synthesizer and a speaker, which has a drawback of high cost.

In addition, the apparatus requires processing related to language and voice, which makes the processing complicated and increases the capacity of the program.

In addition, there is a problem that a noise using voice becomes a problem, a hendphone is required to eliminate the noise, and it takes time and effort to put it on.

On the other hand, in the method using morpheme, syntactic and semantic analysis, the function of proofreading the mistake of proper nouns has not been reported, and as before, the proper noun portion must be pushed out from the sentence and proofreading must be performed after confirmation. However, there is a drawback that the worker is burdened and time is required.

In addition, as a problem common to both formulas, proper nouns often have a large meaning in a document, but the function of extracting the proper noun part and performing confirmation has not been reported. When checking only the part, there is a drawback that the work efficiency is reduced.

The method of the present invention extracts a part related to a proper noun from a sentence by morphological complaint analysis, displays it in a form that allows easy confirmation of the proper noun, and enables a user of the device to easily proofread. It tries to solve the problem.

<Means for Solving Problems> The present invention provides means for inputting / editing Japanese, means for storing the input Japanese, dictionary means for storing a dictionary, and grammar means for storing grammar. And means for extracting a character / symbol string to be proofread from the input Japanese, and morphological analysis of candidate characters / symbol strings corresponding to the input Japanese by the dictionary means and the grammar means. The morpheme portion related to the proper noun is extracted, and the detecting means for detecting the proper noun and the responding to the detecting means are displayed in reverse video so that only the proper noun can be distinguished from other sentence parts. And a means for correcting a candidate character / symbol string to be proofread by the proper noun displayed by the display means, if necessary. -It is a calibration support device.

<Operation> From the Japanese input to the computer by the input means, morpheme analysis is performed to extract the morpheme part related to proper nouns and display it so as to distinguish it from other sentence parts. It works to make it easier to confirm the correctness of the part related to the proper noun.

When you point the extruded part with a pointing device such as a cursor, create a KWIC that uses the proper noun part as a key and display it in another place to further confirm the correctness of the part related to the proper noun. Acts to facilitate.

If there is more than one proper noun part in the sentence, use the information that distinguishes the proper noun part from other parts of the sentence, and add a function that allows you to specify only the extracted part. Further, the function of facilitating confirmation of correctness of the proper noun portion is provided.

In the above confirmation work, if it is confirmed that there is an error in the extracted part, it corrects the error part in the original sentence or KWIC using the input editing function, and acts to proofread the sentence .

Using the checking and correcting functions as described above, the proper noun portion is efficiently proofread, and the above-mentioned problems are improved.

<Example> The present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram of a Japanese sentence proofreading device according to the present invention.

In the figure, 1 is a means for inputting / editing a Japanese character string.

Reference numeral 2 is a means for storing the Japanese character string input by the input means. A keyboard is usually used as the input means, but it is also possible to substitute an external storage means for storing the input Japanese character string such as a floppy disk or a magnetic tape instead of sequentially inputting. That is, there may be a configuration in which 1 is omitted.

3 is a means for storing a dictionary for morphological analysis of the Japanese character / symbol strings accumulated in 2 above.

4 is used for the same purpose and is a means for storing grammar and other dictionaries.

Reference numeral 5 is a control means for extracting a proper noun portion from the character string stored in 2, storing an intermediate result, and controlling display. The control means includes means for storing the result obtained by the control.

Reference numeral 6 is a display means such as a CRT for displaying an input character string, a result of collation, a character string to be calibrated, KWIC and the like.

Reference numeral 7 is a proofreading means for correctly reflecting the correction result in the original text when there is an error in the KWIC displayed by 6.

FIG. 2 shows an example of the character string input by 1, and shows the state displayed on 6 by the control means 5. This sentence is 8. In addition, this sentence is incorrect, and it is correct that it is "this is the invention of Kugara et al."
That is, the proper noun "space" is incorrect, and "void" is correct. The sentence after being proofread correctly is set to 9 in distinction from 8.

FIG. 3 shows an example of a display according to the present invention. This figure shows an example of a result in which the proper noun portion 10 is extracted from the sentence 8 and displayed separately from the others.

FIG. 4 is another example of the display of the present invention using KWIC, showing an example of a state in which a pointing device such as a cursor is placed on the proper noun. In the figure, the part corresponding to the keyword in KWIC is 11, and the whole KWIC is 12.

FIG. 5 shows an example of the contents of the word dictionary. Reference numeral 13 is a headline, 14 is information indicating a distinction between an independent word or an adjunct word, and 15 is information about a part of speech or a classification.

FIG. 6 shows an example of grammar. 16 is the condition of the clause. [] Indicates that the element in it is an indispensable condition for the clause to be established. Other elements may be omitted.

FIG. 7 shows information for assembling morphemes to form a phrase. Reference numeral 17 is a preceding morpheme, 18 is a succeeding morpheme, and 19 is information indicating whether or not the character string connecting these morphemes satisfies the ending condition of the phrase.

FIG. 8 shows information for extracting proper nouns and displaying them so as to distinguish them from others. 20 is a character, 21 is the position of the line where the character exists in the display screen, and 22 is the position of the column. Further, 23 is information for distinguishing whether or not it is a proper noun. 24 describes the start and end points of proper nouns, KW
This information is used when creating ICs. It may be anything as long as it can be distinguished from the others, but in order to make the explanation easy to understand, the start point of the proper noun is set to S and the end point is set to E.

FIG. 9 shows that the specific position 25 of the display device can be described by the numbers of rows and columns of M and N.

FIG. 10 is a schematic flow chart of the present invention. The example sentence in Figure 2,
8 will be described in detail below.

First, the sentence “This is the invention of Mr. Shira et al.” Input by the input / editing means 1 is accumulated in the input storage device 2. 26 is this processing block.

Each input character string is collated with a word dictionary and grammar to determine morpheme elements. 27 is a processing block for word dictionary matching. Here, as a result of the word dictionary search, “kore” is searched first, and it is found that it is a pronoun. Next, "ha" is searched, and it is understood that it is a particle and can be connected to a pronoun, and that the ending condition of the phrase is also satisfied from FIG. It is a processing block that matches the grammar with conditions for clause establishment, conditions for connection between morphemes, conditions for ending phrases, and the like. This is 28.

As a result, there are cases where both the dictionary and the grammar are successfully matched, and cases where they are not. In the latter case, error processing 35 is performed, and the processing for the next character is started.

In process 28, using FIG. 7, it is checked whether or not the connection between the element of the preceding morpheme and the element of the subsequent morpheme. For example, in the case of the pronoun “kore”, it can be seen that this alone does not satisfy the end of the phrase, and the character string is connected to “kowa” to form the phrase. Next, the process of "space" is performed. According to the word dictionary, there are cases where the space is a noun with independent words and a proper noun representing Mr.

In such a case, it is necessary to select either one. In some cases, there is no contradiction in the morphological analysis, which is accumulated in the stack as a candidate, the next morpheme is analyzed, and a proper one can be selected from the candidates in the connection relation between the morphemes. For example, the noun “space” and the suffix “Mr.” cannot be connected, but the proper noun “space” and the suffix “Mr.” can be connected. In such a case, the longest match method is known, the number of clauses is reduced, and the candidate with the longest character string is selected. As a result, "Space Mr." associated with the proper name is selected. By repeating the same processing, "Space Mr. et al." Is cut out as one clause. The processing block that composes this clause is 29.

The collating order of the codes of the input character string is not fixed in particular, and any character does not affect the present invention. Here, we consider the order of reading ordinary texts in the order of "This is a space ...".

By 27, 28, 29, "space ..."
The symbol S is attached to 24, and the letter 24 of the character "ma" is attached to E, and it can be seen that "space" is one proper noun.

Reference numeral 30 is a processing block for determining the end of the morphological analysis processing. If there is a subsequent character, that character is called into the control device, set on the stack, and prepared for morphological analysis. This processing block is 31. This process is repeated until there are no subsequent characters.

Reference numeral 32 is a processing block for displaying the proper noun in the original sentence separately from the others based on the result of the stack in which the information shown in FIG.

33 is another method for extracting and displaying proper nouns. That is, it is a processing block that generates a KWIC using a proper noun as a key.

Reference numeral 34 is a processing block that confirms proper nouns and corrects / edits any mistakes. Among them, the proper noun "space"
Is corrected as "Kuga".

After the confirmation is completed, a series of operations can be completed.

Next, examples of implementation means other than the above description will be described.

In FIG. 8, the information describing the distinction between proper nouns and other characters and the information indicating the start and end of proper nouns are distinguished 23, 24
Although it was saved in, it is possible to use a shared configuration.

Also, in the above explanation, after extracting proper nouns in the original sentence, KW
I made IC, but don't worry about this order. That is, only the proper noun part in the original sentence is extracted and KWIC is not generated or displayed, or KWIC is not displayed.
There may be a method of generating proper names and then extracting proper nouns from the original sentence. The extracted proper nouns may be displayed in a uniform manner, or may be displayed only in a portion designated by a cursor or the like. Whichever method is adopted, the present invention is not affected.

In addition, proper nouns can be calibrated directly by using the extracted proper nouns in the original text, or by using keywords in KWIC and incorporating the results in the original text. Even if the above is adopted, the present invention is not affected.

<Effects of the Invention> An effect of the present invention is that proper nouns have an important meaning, and only proper nouns can be extracted from a document to facilitate confirmation and proofreading.

Further, according to the present invention, the proper noun portion is reliably extracted, which is effective in improving the accuracy of calibration.

Further, since it is possible to omit the work of extracting the proper noun portion from the original sentence, it is effective in reducing the processing time of the proofreading and reducing the mental burden and fatigue of the person who performs the proofreading.

Further, there is an advantage that no special device such as a voice reading device, a speaker, a voice synthesizing device is required.

In addition, there is an advantage that a special device such as a phonophone is not required to be attached, and the labor and the purchase of the device can be omitted.

In addition, complicated processing such as language processing and voice processing is unnecessary, and it is effective in terms of processing program creation, maintenance, and storage capacity.

Further, since it can be executed by a simple process, it is effective in that the processing speed is high.

In addition, since it does not include a noise-causing element such as voice, it is effective in that the configuration and the work can be performed without considering the influence on the surroundings.

[Brief description of drawings]

FIG. 1 is a block diagram of the device of the present invention, FIG. 2 is a diagram showing an example of a character string input by input means, FIG. 3 is a diagram showing an example of display, and FIG. 4 is a display using a KWIC. , FIG. 5 is a diagram showing an example of the contents of a word dictionary, FIG. 6 is a diagram showing an example of a grammar, FIG. 7 is a diagram showing an example of a clause establishment condition,
FIG. 8 is a diagram showing an example of proper noun extraction stack information, and FIG.
FIG. 10 is a diagram showing an example of a method of designating a specific point of a display device, and FIG. 10 is a schematic flow chart of the present invention. 1 ... input / editing means 2 ... input character storage means 3 ... dictionary storage means 4 ... grammar storage means 5 ... control means 6 ... display means 7 ... proofreading means 8 ... example sentences including errors 9 …… Correct example sentence 10 …… Part of proper noun in the sentence 11 …… Keyword in KWIC 12 …… Entire KWIC 13 …… Heading in word dictionary 14 …… Independent word in word dictionary, auxiliary word distinction information 15… … Part of speech in word dictionary, distinction information 16 …… Branch formation condition 17 …… Preceding morpheme 18 …… Succeeding morpheme 19 …… Phrase ending condition 20 …… Character stack for screen display 21 …… Line position stack for screen display 22 …… Column position stack for screen display 23 …… Proper noun decision stack for screen display 24 …… Proper noun decision information stack 25 …… Specific position on display 26 …… Input character string accumulation processing block 27… … Word dictionary matching processing block 28 …… Grammar matching processing block C 29 ... Phrase decision matching processing block 30 ... Morphological analysis end processing block 31 ... Next character processing block 32 ... Proper noun extraction and display processing block 33 ... Proper noun KWIC display processing block 34 ... Proofreading processing block 35 ...... Error processing block

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Taro Morishita 22-22 Nagaike-cho, Abeno-ku, Osaka City, Osaka Prefecture Sharp Corporation (72) Nobuo Nakamura 22-22, Nagaike-cho, Abeno-ku, Osaka City, Osaka Prefecture Incorporated (56) Reference JP 55-43627 (JP, A) JP 60-123960 (JP, A) JP 63-94365 (JP, A)

Claims (1)

[Claims] 1. A means for inputting / editing Japanese, a means for storing the input Japanese, a dictionary for storing a dictionary, a grammar for storing a grammar, and a dictionary for storing the input Japanese. A means for extracting a character / symbol string to be proofed from the inside, a morphological analysis of a candidate character / symbol string corresponding to the input Japanese by the dictionary means and the grammar means,
Detecting means for extracting a part of a morpheme related to a proper noun and detecting the proper noun, and means for responding to the detecting means and displaying only the proper noun in reverse video so as to be distinguished from other sentence parts. A document creation / proofreading support, characterized in that, when there is a candidate character / symbol string to be proofread by the proper noun displayed by the display means, the candidate character / symbol string is corrected. apparatus.