JPH0746359B2 - Japanese sentence processing method - Google Patents

Description

[Detailed Description of the Invention] [Summary] Regarding a Japanese sentence revision method that points out redundant expressions and automatically corrects them into appropriate expressions, in addition to pointing out unnecessary type expressions and redundant expressions in the document, In order to enable automatic correction to an appropriate expression, in a Japanese sentence processing method that points out redundant expressions in Japanese sentences using a revision processing system and automatically corrects them into appropriate expressions, Japanese lexicals that store redundant grammatical information indicating that it is redundant and its correct terms and morphemes that search the Japanese dictionary and analyze the input sentence into morphemes and are redundant Based on the morphological analysis unit that takes in the grammatical information indicating that it is redundant from the word dictionary and its correct term, and points out the redundant expression based on the processing result of the morphological analysis unit and uses the correct term. And a redundant expression check unit for correcting the input sentence.

[Industrial application field]

The present invention relates to a Japanese sentence processing method, and more particularly, to a Japanese sentence revision method that points out redundant expressions and automatically corrects them into appropriate expressions.

[Conventional technology]

There are many opportunities to create sentences in Japanese, but it is difficult to create appropriate sentences. For that purpose, a revision support system has been developed. The revision support system is composed of a Japanese dictionary, a Japanese grammar dictionary, and revision knowledge, and it corrects typographical errors by reviewing the created sentences. Below is an example of automatic correction of inappropriate sentences by the revision support system.

Example → Automatic correction ・ Grammar gradually gradually → Gradual ・ Inappropriate term back Japan → Sea of Japan side ・ Punctuation mark ・ Movie seen yesterday → Movie seen yesterday ・ Corresponding to parenthesis ・ Reading this book (released in September) ・ Style yesterday closed I was late, so ...
… Delayed ・ Duplicate horse was dropped ・ Footless sentence I am correct in this paper ・ Don't work like he does ・ Do not work like him ・ Long string Thank you very much ・ If there is ambiguous damage etc. ・ Language preferred Whether you like it or not →
I don't like it-Double denial doesn't mean I can't make it in time → Make it in time-I can see the effect of misuse → I can see the effect-Long sentence He …………, ……, ………… →
He did ...・ Discontinue continuous use Go to Tokyo and investigate → Go to Tokyo and investigate ・ Multiple language machine translation → Machine translation ・ Unregistered word Chongqing oil field visited ・ Translated by ATLAS omitted ・ Responding tomorrow movie ・The co-occurrence translation rate increases → increases [Problems to be solved by the invention] When a Japanese sentence is created, the meaning of the sentence becomes unclear due to the use of unnecessary fixed expressions and redundant expressions. ， The text may be difficult to read.

It is difficult to remove useless fixed expressions and redundant expressions from the text even if you are careful while writing the text while thinking and when re-reading the text you created.

In particular, expressions such as “about”, “what is ~”, and “to be able to” are easy to use, so it is easy to add them even if they are not necessary, but they are easy to understand and clear in Japan. In order to create a word sentence, such expressions must be excluded.

The conventional revision support system has a problem that it is not possible to point out the useless fixed expressions and redundant expressions as described above.

An object of the present invention is to provide a Japanese sentence processing method capable of pointing out useless fixed expressions and redundant expressions in Japanese sentences and automatically correcting them into appropriate expressions.

[Means for Solving the Problems]

In order to achieve the above-mentioned object, the Japanese sentence processing method of the present invention uses a revision processing system to point out a redundant expression of a Japanese sentence and automatically corrects it into an appropriate expression. Japanese lexicals that store redundant grammatical information indicating that it is redundant and its correct terms and morphemes that search the Japanese dictionary and analyze the input sentence into morphemes and are redundant A morphological analysis unit that takes in grammatical information indicating that it is redundant from a word dictionary and its correct terms, and a redundant expression that points out redundant expressions and corrects the input sentence using correct terms based on the processing result of the morphological analysis unit. A check unit is provided.

FIG. 1 is an explanatory view of the principle of the present invention.

In FIG. 1, 1 is a Japanese dictionary, 2 is a morphological analysis unit, and 3 is a redundant expression checking unit.

The Japanese dictionary 1 stores, in addition to attribute information such as semantic attributes and grammatical attributes, grammatical information indicating that the redundant expression is redundant as a headword and its correct term.

The morpheme analysis unit 2 retrieves the Japanese dictionary 1 to analyze the input sentence into morphemes, and fetches from the Japanese dictionary 1 grammatical information indicating that the morpheme is redundant and its correct term.

The redundant expression checking unit 3 points out a redundant expression based on the processing result of the morphological analysis unit 2 and corrects the input sentence using a correct term.

[Action]

 The operation of the present invention will be described using a specific example.

Now, it is assumed that the sentence "For new additions, it can be omitted if the guidebook is revised."

The input sentence is morphologically analyzed by searching the Japanese dictionary 1 in the morphological analysis unit 2. That is,
“However,” “new,” “added,” “minute,” “about,” “guidance,” “a,” “revision,” “approved,” “case,” “omitted,” “can, can” It Each morpheme "however", "new", "addition", "minute", "about", "handbook", "wa",
The "revision", "do it", "case", "omit" and "you can" have attribute information such as semantic attributes and grammatical attributes stored in the Japanese dictionary 1. It

In particular, the morphemes “however”, “about”, “do you did”, and “do you can” are stored in the Japanese dictionary 1 because grammatical information indicating that they are redundant is stored. The morpheme also has grammatical information indicating that it is redundant and correct terms.

The pattern is shown below.

(Morpheme) (Grammar information, correct term) (However) = (Redundant, but) (About) = (Redundant, is) (Yes) = (Redundant, Yes) (Can) = ( Then, the redundant expression check unit 3 points out the redundant morpheme, corrects the input sentence using the correct term, and outputs the result as a message. That is, redundant expression: However → For redundant expression: → is made redundant expression: is done → Redundant expression: can be done → Can be modified sentence However, new addition sentence, when the guidebook is revised ， Can be omitted.

The message is output as follows.

〔Example〕

 FIG. 2 is a block diagram of an embodiment of the present invention.

In FIG. 2, 21 is an input / output device, 22 is an input file, 23
Is the revamp control unit, 23a is the function selection unit, 23b is the revamp control function, and 23c
Is a correction control function, 24 is a refining processing unit, 24a is a refining engine, 24
b is a morphological analysis unit, 24c is a syntax analysis unit, 24d is a redundant expression check unit, 25 is a Japanese dictionary, 26 is a Japanese grammar dictionary, 27 is an output file, and 28 is a refinement result processing unit.

The input / output device 21 inputs a Japanese sentence to be refined,
This is a device that retrieves the results of refinement. It may consist of a batch device, an interactive device, or both.

The input file 22 stores Japanese sentences to be refined.

The elaboration control unit 23 analyzes a command relating to elaboration or revision and controls the corresponding processing unit, and includes a function selection unit 23a, a revision control function 23b, and a revision control function 23c.

The function selection unit 23a follows the instruction from the input / output device 21.
Select the function of revision or correction.

The elaboration control function 23b activates the elaboration processing unit 24 in accordance with the elaboration control command.

The modification control function 23c activates the revision result processing unit 28 according to the modification control command.

The refinement processing unit 24 includes a refinement engine 24a, a morpheme analysis unit 24b, a syntax analysis unit 24c, and a redundant expression check unit 24d.

The revision engine 24a controls the morpheme analysis unit 24b, the syntax analysis unit 24c, and the redundant expression check unit 24d.

The morpheme analysis unit 24b searches the Japanese dictionary 25 and analyzes the input sentence into morphemes. In addition, the redundant morphemes are provided with grammatical information indicating that they are redundant and correct words stored in the Japanese dictionary 25.

The syntactic analysis unit 24c uses the grammatical dictionary 26 based on the information obtained by the morphemes analyzed by the morphological analysis unit 24b.
To find the structure of the input sentence.

The redundant expression checking unit 24d points out redundant expressions and corrects the input sentence using correct terms based on the processing result of the syntax analysis unit 24c.

The Japanese dictionary 25 stores Japanese word expressions and attribute information such as semantic attributes and grammatical attributes that the Japanese dictionary 25 has. In the case of the present embodiment, particularly redundant notation is used as a headword, and grammatical information indicating that it is redundant and its correct term are stored.

The Japanese grammar dictionary 26 stores Japanese grammar, that is, syntactic rules.

To the output file 27, the result of rectification of the input sentence rectified by the rectification processing unit 24 is output.

The revision result processing unit 28 corrects the input sentence based on the revision result data stored in the output file 27 according to an instruction from the revision control function 23c, and the result is passed through the revision control unit 23 to the input / output device 21. Output to.

The operation of this embodiment will be described below.

In the case of interactive revision, the input / output device 21 is composed of a keyboard and a display, Japanese sentences to be revised are entered one by one from the keyboard, and the revision result is displayed on the display screen. In the case of collective revision, I / O device 21
Is composed of a keyboard and a printer, the Japanese sentences to be elaborated are stored as sentences in the input file 22, and the elaboration result is printed by the printer.

In both cases of interactive and collective revision, after inputting a sentence to be revised, the input / output device 21 gives an instruction to the function selection section 23a of the revision control section 23, and first, the revision control function 23b.
To start.

The elaboration control function 23b transfers the input sentence data to the elaboration engine 24a of the elaboration processing unit 24.

The refining engine 24a activates the morphological analysis unit 24b and transfers the input sentence data.

The morpheme analysis unit 24b searches the Japanese dictionary 25 and analyzes the input sentence into morphemes. At this time, if the analyzed morpheme is a redundant expression, the morpheme has grammatical information indicating that it is redundant and a correct term.

When the morphological analysis is completed, the revision engine 24a activates the syntax analysis unit 24c.

The syntactic analysis unit 24c obtains the structure of the input sentence by searching the Japanese grammar dictionary 26 based on the morphemes analyzed by the morphological analysis unit 24b.

When the syntax analysis is completed, the revision engine 24a activates the redundant expression check unit 24d.

The redundant expression check unit 24d corrects the input sentence by pointing out the redundant expression and using the correct term. The resulting data is output to the output file 27. It is also output to the input / output device 21.

Next, from the input / output device 21 to the function selection unit 23 of the revision control unit 23.
The correction control function 23c is activated via a.

The modification control function 23c activates the modification result processing unit 28 to modify the modification result data stored in the output file 27. For example, if there is a redundant expression in the input sentence, correct it with the correct term.

The processing result of the selection result processing unit 28 is output to the input / output device 21 through the selection control unit 23.

FIG. 3 is a processing flow chart of the refining engine.

The operation of the refining engine will be described in more detail with reference to FIG.

Capture the Japanese sentence to be elaborated.

Start the morphological analysis unit.

The morpheme analysis unit searches the Japanese dictionary and analyzes the input sentence into morphemes.

Determine if there are redundant expressions in morphemes.
If YES, execute the next process, and if NO, end.

The morpheme, which is a redundant expression, has grammatical information indicating that it is redundant and its correct term from the Japanese dictionary.

Start the redundant expression check unit.

The redundant expression check unit points out a redundant expression based on the processing result of the morphological analysis unit.

The redundant expression checking unit then corrects the input sentence using the correct term.

Outputs a message indicating that it is redundant and a correction sentence.

Here, an example of the redundant expression check will be described.

Now, "~, It is necessary to notify you of the change. However, for new additions, if the guidebook is revised, it can be omitted. In the case of revision, the following points Must pay attention to. "
It is assumed that the sentence is input.

When this input sentence is analyzed into morphemes, "~""change"
"No", "Notification", "O", "Need", "However", "New Addition", "About", "Guide"
"", "Revision", "tried", "case", "omission"
"Can", "Revise", "If", "About", "Below", "Of", "Point", "To", "Pay attention"
It becomes "must have".

On the other hand, the Japanese dictionary stores the following redundant expressions.

(Heading) (grammar information, correct term) (must be done) = (redundant, must be) (however) = (redundant, but) (for) = (redundant, is) ( ) = (Redundant, Yes) (Can be) = (Redundant, Yes) (Revise) = (Redundant, Revision) (Pay attention) = (Redundant, Be careful) Therefore, redundant expression A morpheme that has information that is redundant and its correct term.

Next, redundant expressions are pointed out based on the result of the morphological analysis. An example is shown below.

P0001 S002: It is necessary to notify the change.

7th character □ Redundant expression: Need to be done → Need to be done S003: However, if the manual is revised, the new addition can be omitted.

1st character □ Redundant expression: However, → 13th character □ Redundant expression: About → 25th character □ Redundant expression: Enabled → 37th character □ Redundant expression: Can be → Yes S004 When paying attention to the following points, you need to pay attention to the following points.

1st character □ Redundant expression: revise → revise 8th character □ Redundant expression: for → is 19th character □ Redundant expression: pay attention → be careful As shown above, the redundant expression and correct term are Appears on the display.

To correct the check result automatically, move the cursor to the position of □ displayed on the display, and
You can change to.

If you want to manually modify the expression to your liking, move the cursor to the place of "→", overwrite the character string you want to modify on the one output by the system, and change the □ mark to ■.

Below is the revised sentence.

"~, You need to notify me of the change. However, the new addition can be omitted when the guidebook is revised. When revising, the following points must be noted."
(The underline indicates the corrected part.) [Advantage of the Invention] According to the present invention, when there is a useless fixed expression or redundant expression in a Japanese sentence, it can be pointed out and appropriate. It will be possible to automatically correct the expression.

[Brief description of drawings]

FIG. 1 is a diagram illustrating the principle of the present invention, FIG. 2 is a configuration diagram of an embodiment of the present invention, and FIG. 3 is a process flow chart of a refining engine. In Fig. 1: 1: Japanese dictionary 2: Morphological analyzer 3: Redundant expression checker

Claims (1)

[Claims] 1. In a Japanese sentence processing method for pointing out a redundant expression in a Japanese sentence using a revision processing system and automatically correcting it into an appropriate expression, it is redundant that the redundant expression is a headword. Japanese dictionary that stores grammatical information and its correct terms (1)
And a morpheme analysis unit that retrieves the grammatical information indicating that it is redundant from the Japanese dictionary (1) and correct correct terms for the morpheme that is redundant while searching the Japanese dictionary (1) and analyzing the input sentence into morphemes ( 2) and a redundant expression check unit (3) for pointing out redundant expressions based on the processing result of the morphological analysis unit (2) and correcting the input sentence using the correct term. Japanese sentence processing method.