JP2006172231A - Document proofreading method and document proofreading apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To develop a proofreading system which can rapidly and accurately search a dictionary for proofreading for a correction object character string registered by a proofreader beforehand, and can display the character string and extend the dictionary. <P>SOLUTION: With a character string detected from proofreading object text by a proofreading object character string detection part 2 set as a search key, a display part displays a character string which the proofreader has recognized as the one to be corrected and has registered in the dictionary 3 for proofreading beforehand, and if a correction history search part 4 has searched for the presence/absence of registration thereof to find the absence of registration, displays as such. If correction is needed, an alternative candidate therefor is retrieved and is also displayed. If the proofreader inputs a character string for the alternative candidate or correction, an alternative candidate decision part 5 decides the character string as the alternative candidate. A sentence update part 6 substitutes the alternative candidate for proofreading object expression. A correction history registration part 12 associates the proofreading object expression with the alternative candidate and registers them in the dictionary 3 for proofreading. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a document proofreading method and a proofreading apparatus (hereinafter, may be abbreviated as “the proofing system of the present invention”). The present invention relates to an apparatus capable of displaying (preferred expression), correcting the error in accordance with the instructions of the proofreader, further displaying, storing and outputting the correction result, and a calibration method for performing the above analysis and the like.

A word processor (abbreviated as “word processor”) having a function of correcting a proofreading target expression such as an error in a Japanese sentence, an inconsistency in a Japanese syllabary, and a forbidden term has been commercially available since about 1988. Some word processing software that has a function for proofreading these proofreading expressions only provides detection of proofreading expressions, and simply presents alternative candidates to replace (replace) the detected proofreading expressions. However, there are those that select any one of a plurality of alternative candidates presented as an alternative expression and replace the proofreading target expression in the original text with the alternative expression.
For example, Patent Documents 1 to 8 propose a system that is installed (installed) in a computer and used for proofreading sentences. As a typical example, one disclosed in Patent Document 1 can be cited.

Japanese Patent Laid-Open No. 05-120275 JP 05-135095 A JP 05-298301 A JP 05-298302 A JP 05-324647 A Japanese Patent Laid-Open No. 06-203609 Japanese Patent Application Laid-Open No. 07-141362 Japanese Patent Laid-Open No. 09-114831

The document creation support device, the push support device, the proofreading device, or the execution method thereof disclosed in each of the above-mentioned published patent publications (documents) are conceptually from an input device, a control device, a display device, a storage device, and others. It is an organized combination.
However, in each of the above-described inventions, the concept of the system is merely displayed in the form of a flowchart, and an embodiment for supplementing it is not detailed in detail. That is, in these examples, although the example sentence to be proofread is presented, it should be said that it is a simple typographical error such as a typographical error or omission, or a complex typographical error such as a phrase or short sentence change, insertion or deletion, etc. It is still an abstract disclosure of how things are rectified.
Also, departments or organizations responsible for document submission, such as patent offices and law firms, often submit manuscripts or previous drafts from the requester when the deadline for submission of documents is extremely imminent. Meet. Even in this case, the submitting agency is obliged to prepare a complete document, so it is necessary to proceed with the calibration as efficiently as possible. To that end, no matter how competent the dictionary installed in normal word processing software is, a dictionary that stores expressions that can also be handled in the stored state (Default) is required. The ability that this type of dictionary should have will be described later.

In view of the above circumstances, the present inventor diligently studied a method and apparatus for urgent document proofreading, and completed the present invention.
In other words, the first aspect of the present invention is a document proofreading system, in which one or more character strings registered in advance by a proofreader and recognized by the proofreader as requiring correction are subject to proofreading. The present invention relates to a sentence proofreading method characterized in that it is detected from the whole sentence or a specified part of a sentence, a detected character string is displayed, and a proofreader instructs the character string to be corrected if necessary after confirmation.
According to a second aspect of the present invention, when the proofreader instructs correction of the displayed character string, the proofreader selects an alternative expression from the displayed alternative candidates and causes the displayed character string to be corrected. The present invention relates to the first sentence proofreading method of the present invention.
According to a third aspect of the present invention, when an alternative candidate is not displayed when the proofreader instructs correction of the displayed character string, the proofreader inputs an alternative expression and corrects the displayed character string. The present invention relates to a first or second sentence proofreading method of the present invention.
According to a fourth aspect of the present invention, when the displayed character string meets a predetermined correction condition, the character string is automatically replaced with an alternative expression to correct the displayed character string. The present invention relates to the first or second sentence proofreading method of the present invention.
The fifth aspect of the present invention is an input function unit, a proofreading target expression detecting function unit that detects a proofreading target expression in a proofreading target sentence, and a correction history registration dictionary that includes a proofreading target expression and sentence proofreading. A dictionary in which both the proofreading target expression and the replaced alternative expression are registered in association with each other, and a correction history search function unit based on the correction history registration dictionary. The present invention relates to a sentence proofreading apparatus that is basically composed of a function part for searching for an alternative expression and a display function part using a detected proofreading expression as a search key.
According to a sixth aspect of the present invention, the sentence proofreading apparatus further includes an alternative candidate determination function unit that determines an alternative candidate from an alternative expression registered in the correction history registration dictionary in association with the proofreading expression. The present invention relates to a fifth sentence proofreading apparatus according to the present invention.
According to a seventh aspect of the present invention, the sentence proofreading apparatus further causes the proofreading target character string to be corrected using the alternative expression input by the proofreader as an alternative candidate when the alternative expression is not registered in the correction history registration dictionary. The present invention relates to a fifth sentence proofreading apparatus of the present invention.
According to an eighth aspect of the present invention, the text proofreading apparatus further includes a function unit that automatically replaces and corrects the proofreading target expression and the alternative candidate determined by the alternative candidate determining function unit. The present invention relates to a fifth sentence proofreading apparatus according to the present invention.
According to a ninth aspect of the present invention, the sentence proofreading apparatus further includes a correction history registration function unit that associates the replaced proofreading expression with the alternative candidates and additionally registers them in the correction history registration dictionary. The present invention relates to a sentence proofreading apparatus according to any one of the fifth to eighth aspects of the present invention.
According to a tenth aspect of the present invention, there is provided a selection function unit for selecting whether or not to register with the correction history registration function unit according to an instruction from a proofreader. The present invention relates to a sentence proofing apparatus.
The eleventh aspect of the present invention further includes a selection function section for selecting whether or not to register with the correction history registration function section in accordance with an instruction from the proofreader. It relates to any proofreading device.
A twelfth aspect of the present invention is a mode switching function section, comprising a mode switching function section that is automatically performed when the replacement function section is instructed by a proofreader. To a text proofreading apparatus according to any one of.
According to a thirteenth aspect of the present invention, the mode switching function section includes a function section for switching modes so that the mode is automatically activated when the replacement function section is instructed by a proofreader. It relates to a twelfth sentence proofreading apparatus.

According to the present invention, it is possible to proofread a sentence in a short time by a simple operation. Moreover, it is possible to unify terms within the same document and between a plurality of documents.
The present invention will be described in detail below.

The present invention is an improvement of the above-described prior art disclosure, which is not yet clear, and particularly relates to an improvement of a dictionary (proofreading dictionary) used for document proofreading.
Specifically, the present invention uses a proofreading dictionary (correction history registration dictionary) that can handle proofreading expressions that cannot be handled by a dictionary installed in normal word processing software, a document proofreading system equipped with the dictionary, and the system. The present invention relates to a document proofreading method.

A text proofing apparatus and a document proofing method will be described below with reference to the drawings.
FIG. 3 is a block diagram or flowchart showing the configuration of a conventional text proofreading apparatus (FIG. 3 of Japanese Patent Laid-Open No. 05-120275) with blocks and lines connecting them. In FIG. 3, reference numeral 21 denotes a sentence storage unit, which is a functional unit that stores Japanese sentences input from the input unit 26. In FIG. 3, reference numeral 22 denotes a proofreading object expression detection function unit that analyzes Japanese sentences based on a Japanese analysis dictionary (not shown) and detects errors in sending kana, prohibited words, etc. as proofreading objects. . Reference numeral 23 denotes an alternative expression dictionary memory, which is a functional unit that stores proofreading target expressions and alternative candidates as appropriate expressions corresponding thereto. Reference numeral 24 denotes an alternative expression display control unit, which is a function unit that searches the alternative expression dictionary memory 23 for alternative candidates stored in correspondence with the proofreading expression and displays them on the display unit 27. Reference numeral 25 denotes a correction / correction control unit, which allows the proofreader (operator) to select an alternative candidate used for replacement from among the alternative candidates displayed on the display unit 27, and the proofreading target expression in the original text by the selected alternative candidate. It is a functional part to be replaced. As described above, reference numeral 26 denotes an input unit, which is a functional unit that accepts input or instructions of Japanese text by a proofreader. Reference numeral 27 denotes a display unit as described above, which is a functional unit that displays the content of the sentence proofreading expression, the detected proofreading target expression, alternative candidates, and the like.
The operation of the sentence proofing apparatus as described above will be described below. First, when the proofreader issues an instruction from the input unit 25 to detect prohibited words as proofreading expressions from the Japanese sentences stored in the sentence update unit, the proofreading expression detection unit 22 analyzes the Japanese sentences. , Detect inappropriate terms from them. For example, if “blind person” is detected as a prohibited word, and the alternative expression display control unit 24 detects an alternative expression “visually impaired” for “blind person”, the display unit 27 displays the proofreading expression and alternative candidates. indicate.
Next, it is assumed that the proofreader selects “visually impaired” from the alternative candidates as an alternative expression and inputs a replacement instruction. Correspondingly, there remains a need for the input unit 25 to instruct the alternative candidate selection instruction from the proofreader, the selection of the alternative candidate corresponding to the proofreading expression in the document, and the replacement of the proofreading expression by the selected candidate. As a result, in the case of proofreading extremely long sentences, the number of operations is increased, and as a result, there is a problem that a situation where a long time is required for the work is often encountered.

The `` proofreading dictionary '' that plays the most important role in the present invention is different from any conventional electronic coding dictionary, and the proofreader himself / herself puts the lexical phrase (character string) recognized by the proofreader himself into the dictionary. It can be said that it was created by registration (storage) and was not sold or transferred anywhere in the world. It should be noted here that the proofreading dictionary of the present invention has the same character string as the registered character string regardless of whether or not the character string registered in it represents common sense. If it exists in the text to be proofread, the mission is to detect it accurately. Here, “it has nothing to do with whether or not the character string registered in the proofreading dictionary according to the present invention expresses reasonable content” is “character string to be corrected” exemplified in the second embodiment. If you look at it, you will be immediately convinced.
However, the meaning that the proofreading dictionary of the present invention was “not sold or assigned anywhere in the world” means that the first proofreading dictionary was “not sold or assigned anywhere in the world”. This is not to say the circumstances after many dictionaries of this kind have already appeared in the world. That is, after the first proofreading dictionary of the present invention stores a certain amount of “character string to be corrected”, the proofreading dictionary that is copied and stored can appear in different places in the world. Of course.

The present invention is intended to solve the problems associated with various document proofing apparatuses and methods provided so far. Specifically, the present invention describes various deficiencies inherent in a document to be submitted, that is, fonts, as “necessity for improvement” that is difficult to deal with with a kanji conversion dictionary or document proofreading software installed in a conventional word processing software. Inconsistencies, typographical errors, omissions, as well as inconsistent use of terminology, prohibited words, inconsistencies found in connection relations or changes in the end of the idiom, improper character or line feed, or style (number of characters per line And the function of pointing out inconsistencies in the number of lines on one page, top / bottom / left / right margins, etc.).
This proofreading of documents immediately before submission is often required to be completed within a very limited time. Therefore, in this work, various revision requirements contained in the document can be searched quickly and without omission, the revision alternative candidates are presented as necessary, and alternative expressions selected from the alternative candidates are displayed by the proofreader. Replace according to instructions (input). If an appropriate alternative expression is not found in the presented alternative candidates, or if no alternative candidate is presented, the alternative expression input from the proofreader is replaced according to the proofreader's instruction.
In each of the above prior inventions, since the configuration of the proofreading dictionary has not been specifically disclosed, it is apparent from the details how the dictionary can demonstrate the document proofreading ability. It was difficult to see. Also, these proofreading dictionaries are different from those registered by the proofreader that need to be corrected like the configuration dictionary of the present invention, so that correct terms are registered and those that do not apply to correct terms are detected. The calibration was done by the method of doing.
The proofreading dictionary, which is the most important function part constituting the document proofreading method and document proofreading apparatus of the present invention, detects various proofreading expressions (necessary proofreading or proofreading preferable description) that are actually encountered as much as possible and detects them. If there is an alternative candidate, it is displayed. If there is no alternative candidate or there is no alternative preferred expression in the alternative candidate, the current expression is replaced with the current expression instead of the current description according to the input of the proofreader. It is a functional part to be replaced with. For that purpose, it is preferable that the document proofreading dictionary of the present invention “stores (registers) various proofreading expressions (necessary proofreading or proofreading suitable description) that are actually encountered” as much as possible. ” Therefore, the document proofreading method and the document proofreading device of the present invention are formed so that the proofreading target expressions encountered at the time of actual text proofreading can be stored in the proofreading dictionary each time or many times at once. As the accumulation of these proofreading object expressions increases, the proofreading accuracy of the document increases, and the past proofreading results can be used for future proofreading.

  In particular, the present invention relates to an improvement of a proofreading dictionary (also known as a “correction history registration dictionary”) that can be said to be the key to proofreading documents. There may be various methods for improving the proofreading dictionary, but in the present invention, as an example of this, as an example of an existing word processing software, a macro program installed in the product name “Ichitaro” of Justsystem Corporation. We propose a method that uses a system (abbreviated as “macro”). In order to use this macro, a method of registering a specific character string group to be searched for in a search key can be mentioned. As a specific method for creating the improved proofreading dictionary of the present invention, a combination of a searched character string and an arbitrary searched character can be adopted. In “Ichitaro V12”, “\?” And a line break in a program are used. In the macro of “Ichitaro V12”, “\ n” is used in “Ichitaro” as a line feed (enter) mark in the searched character string. What should be emphasized about the proofreading dictionary of the present invention is that the proofreading dictionary not only contains normal regular words (words or phrases) in its memory, but rather contains many irregular words or typographical words. . In other words, the fact that the proofreading dictionary of the present invention exerts its power is that it “contains many irregular words or typographical phrases”. In other words, the system (the proofreading dictionary) can be found because the same or approximate expression (if any search target character is included) is pre-stored in the dictionary and exists in the proofreading text. Of course, it is not possible to detect irregular expressions. For example, “useless medium” is an error of “no solvent”, but the system cannot naturally detect this “useless medium”. This is because the system does not store “useless media” in the calibration dictionary. Similarly, “1 general” is an error of “general”. “Looking value” is an error of “Looking value”. It is useful at a later date to register these erroneous expressions (erroneous character strings) in the proofreading dictionary of the present invention every time they are found.

This kind of error is also due to the lack of national knowledge in the creator. Furthermore, there was an example in which “5Oz (ounce)” was erroneously described as “50z” in the specification. This was a typographical error due to the typist misreading the Roman letter “O” as “0”. However, it is impossible to expect all typists to be able to read even abbreviations of foreign units of measure. In any case, these documents should not be submitted in the wrong way. If there is a certain number of this type of error, not just one, it is difficult to correct or eliminate them without omission within a very limited time before the deadline for submission. . To that end, the document proofing system of the present invention is mounted on a computer, and the proofreading operation is performed at high speed without omission.
Therefore, it can be said that the present invention has its usefulness because the proofreading dictionary includes as many “words to be corrected, possibly erroneously written phrases” as in the above example as much as possible. Unlike a normal dictionary, the proofreading dictionary of the present invention is positioned so that a lexical word (character string) that a proofreader recognizes to be corrected in advance is registered in advance. In the above example, “0Z” or “0z” is registered once, and this character string can be found in future document proofreading. In other words, each time the proofreader finds a lexical phrase (character string) that is difficult to be registered (stored) in advance in an ordinary dictionary, the proofreading dictionary of the present invention is subjected to importance determination in relation to the correction target and is important. The determined items are registered with priority (subjective). Therefore, in most cases, the stored phrases (stored character strings) are of a character that makes it difficult to understand what they mean by themselves.

Here, “pre-registered and placed” includes a case of registering and placing immediately before. That is, in the same document, it is possible to play a role of correcting (including term unification) later synonymous words and phrases by reliably registering the first words and phrases (character strings) in the document. Since this type of proofreading dictionary is not found in the preceding documents, it can be said that the proofreading dictionary of the present invention has sufficient specificity even with the above-mentioned feature alone. However, since the proofreading dictionary of the present invention first reaches a certain number of stored phrases, it is naturally possible to transfer the contents to another proofreading dictionary. It is also possible to realize a case where a certain number of lexical phrases (character strings) useful for proofreading are stored in advance.
The proofreading dictionary of the present invention is organized so that not only “character strings” can be registered in advance but also new “character strings” can be registered in the course of use. However, as described above, this “character string” does not need to have a normal meaning or validity. That is, the role of the `` proofreading dictionary '' installed in the proofreading software of the present invention is completely different from the normal `` dictionary '' software installed (stored) in word processing software etc., `` correct lexical (character string) Is not meaningful. On the other hand, the proofreading dictionary installed in the proofreading software of the present invention is a `` character string (lexical phrase) recognized by the proofreader as requiring correction '', i.e., in many cases, the reader itself makes sense. Storing a character string that doubles as a main object has its significance (mission).

As the above-mentioned character string “reading by the proofreader”, an unexpected typographical word that repeats due to the personality (癖) of the drafter or proofreader may occur. Since this type of “character string that doubles as reading” can be different for different proofreaders, there is no inevitable commonality between them. However, even if there is a commonality in a certain range, it is only a coincidence.
When a document proofreading method and apparatus of the present invention are used simultaneously by a plurality of proofreaders, the system of the present invention is mounted on a common server computer (abbreviated as “server”) and each client computer (abbreviated as “client”). )) And can be organized so that each proofreader has independent access to the server from its clients. This client-server system is very practical in many document creation institutions such as patent offices. In this document proofreading system, it is useful that each proofreader can freely access the proofreading dictionary and that the “character string to be corrected” discovered by each writer can be freely written in the proofreading dictionary. is there. However, it is extremely important that a single person in charge of unification is responsible for correcting, organizing, and organizing the registered character strings in this proofreading dictionary without leaving it to each proofreader. Note that the person in charge of unification does not necessarily need to be a single person for the entire agency. When the document creation organization has multiple departments, it is beneficial to have a uniform manager in each department. Although not shown in the present specification, an input function unit can be given as an example of a function unit that is normally attached. The so-called “input function unit” is a function unit that is extremely useful for proofreading a document, and is a function unit that plays a role in electronically encoding a document to be proofread into the system. is there. Therefore, the “input function unit” includes, for example, a keyboard, a mouse, a scanner (necessary to use with OCR software), a magnetic floppy disk drive (FDD), an MO drive (MOD), a DVD drive, a magnetic tape reader, and a magnetic signal. It may be a reader or other computer.

A fifth sentence proofreading apparatus according to the present invention is an apparatus in which the proofreading expression detection function unit detects a proofreading expression in a proofreading sentence based on the correction history registration dictionary. In this detection, the detection operation is registered to the keyboard key and started by key operation. The search is repeated and the search automatically proceeds as the proofreading page advances to the proofreading expression in the proofreading document. It is also possible to calibrate by displaying with a mark or by inverting the black and white. The correction history detection function unit is a function unit in which the proofreading target expression and the replaced alternative expression are registered in association with each other.
The correction history search function unit can be provided as necessary, and based on the correction history registration dictionary, the proofreading target expression detected by the proofreading target expression detection function unit is used as a search key to search for alternative expressions. If the proofreading target expression is registered in the correction history registration dictionary, the alternative candidate determination function unit sets the associated alternative expression as a replacement candidate, while if not registered, the alternative expression input by the proofreader Is an alternative candidate.
The replacement function unit automatically replaces the proofreading target expression with the replacement candidate determined by the replacement candidate determination function unit. The correction history registration function unit is a function unit for additionally registering in the correction history registration dictionary in association with the proofreading target expression that has been replaced. In the seventh sentence proofreading apparatus of the present invention, according to the instruction input by the proofreader, the selection function unit selects whether or not to register with the correction history registration function part of the fifth sentence proofreading apparatus of the present invention.
In the twelfth sentence proofreading apparatus of the present invention, the mode switching is performed so that the replacement function part of the fifth sentence proofreading apparatus of the present invention automatically activates the mode switching function part when instructed by the proofreader. Do.
In the sentence proofreading object expression detection function unit of the present invention, the correction history detection function unit can also record irregular words in a registered state in addition to normal regular words.

Example 1
In the following, the present invention will be described in the form of several embodiments of a text proofing apparatus with reference to the drawings. FIG. 1 is a block diagram showing an example of a sentence proofreading apparatus of the present invention, and the present invention is not limited to this. In FIG.
Reference numeral 1 denotes a sentence storage function unit, which is a function unit (storage device) that holds Japanese sentences input from an input function unit (not shown). This text storage function unit 1 can have any configuration as long as it can store Japanese text such as magneto-optical disk (MO) or floppy magnetic disk (FD), CD ± R (W) or DVD ± R (W). Even things can be used. The Japanese text may be input from a keyboard or the like, input as a result of speech recognition or character recognition, or transferred from another device via a network or FD. The input function unit may be, for example, a keyboard, an image scanner (requires OCR to encode character input), a card reader, and other code readers.
Reference numeral 2 denotes a proofreading target expression detection function unit (sentence error detection device), which is a function unit that performs a function that can also be referred to as an alias “sentence analysis function unit”. This detection function unit is one or more character strings registered in advance by a proofreader from Japanese sentences stored in the sentence storage function unit 1 based on a Japanese analysis dictionary (not shown). This is a function unit that searches for a proofreading expression such as a character string in which a character string that the proofreader intends to correct is stored in a proofreading dictionary. Here, the expression “pre-registered by the proofreader” means that the proofreader himself consistently performed the operation to memorize (register), and the proofreader switched the system to the automatic mode, The case where the step is automatically performed and the storage operation is stopped when the end of the storage is recognized is also included.
(2.1.) The above-mentioned character string (expression) registered in advance by the proofreader is mainly “unnatural lexical connection (interaction), error in sending kana and inconsistency in sending kana etc. , "Unfixed terms such as prefix or suffix". Also, there are words that are often mistaken in the field that is the subject of the document content, for example, “publication” is “public information”, and “reflux” is “circulation” in the distillation field. In particular, what corresponds to this kind of “unnatural connection between lexical phrases (connection = engagement)” is a lexical phrase (character string) that can generally be understood only by looking at the face. For example, "... is extremely good ~"("Ta" should be "te"), "... become."("Te" should be "ta"), "Think ("Think"), "Coating"("Coating"), and so on. In addition to these, a case where a character string (including symbols) that should be half-width characters is input in full-width characters and vice versa can also be cited. In the same manner as above, when a lexical character to be input in kanji is input in hiragana or katakana, and when a lexical character to be in hiragana is input in kanji or katakana, it is added to the “target expression to be corrected” in the present invention. It is done. For example, “to be different” (should be “different” in kanji, not hiragana).
(2.2.) It should be noted that the “detect proofreading expression previously registered by proofreader” function part of the present invention has any of the following functions provided in normal document creation software (such as word processing software). It does not have. In other words, a function for detecting a proofreading target expression such as a prohibited term (mainly various discriminatory expressions, a discriminatory term, etc.) In addition, it does not have any function such as “line end prohibition”.
(2.3.) Among the above-mentioned specific character strings, those that can only be detected by the detection function of the present invention are “unnatural connection between lexical characters preliminarily registered by a proofreader” and “between lexical characters preliminarily registered by a proofreader. Is an unnatural flow. The document proofreading system of the present invention is organized so that it can be registered each time the "unnatural connection between lexical phrases" encountered during proofreading of sentences. It leads to the result of acquiring the ability to detect without.
Reference numeral 3 denotes a correction history registration dictionary, which is a function unit that can also be called an alias “analysis result storage function unit”. This memory function unit is sufficient to have the commonality with the above-mentioned `` text memory function unit '', for example, by associating the proofreading target expression with the alternative expression replacing the proofreading target expression at the time of proofreading the sentence. It is stored in a rewritable memory. This dictionary basically performs the same function as the “correction history registration function” described later. As a method of associating the proofreading target expression with the replaced alternative expression, for example, the expression itself to be proofread detected in the above 1 and the appearance position thereof are stored and stored in the proofreading dictionary. It is displayed in contrast with the alternative candidate expression. Storage media include HD, CD (for example, CD ± R and CD ± RW) or DVD (any of DVD ± R and DVD ± RW and DVDRAM), magneto-optical disk (MO), or floppy magnetic disk (FD) Any structure that can memorize Japanese sentences such as magnetic tape can be used.
Reference numeral 4 denotes a correction history search function unit, which uses a proofreading expression character string as a search key, and selects an alternative expression from among alternative candidates stored (registered or stored) in the correction history registration dictionary 3 in a form associated with the search key. It is a functional part to search. The premise for this function to work is that the association has already been done more than once. However, the search unit of the present invention can be organized so as not to detect expressions that do not correspond to the stored alternative expressions. This type of expression mainly corresponds to the above-described “unnatural connection between words”. That is, this kind of “target expression to be corrected” is difficult to register (store) in the default (stored) proofreading dictionary, so when this kind of expression is encountered, it is stored in the proofreading dictionary. In principle, it is necessary to input for this purpose. This input can be done automatically or according to the instructions of the proofreader.
Reference numeral 5 denotes an alternative candidate determination function unit. When the calibration target detected by the calibration target detection unit 2 is stored in the correction history registration dictionary 3, the associated alternative expression is set as the alternative candidate. On the other hand, when the detected proofreading target expression is not stored in the correction history registration dictionary 3, the function unit uses an alternative expression input from the alternative expression input function unit 10 described later as an alternative candidate. In addition to the above functions, it is useful that this alternative candidate determination function unit has the following functions. That is, if a new character string exists in the character string constituting the input sentence in the process in which the alternative candidate determination function unit analyzes the input sentence (proofreading sentence) in character string units, Refers to the contents of the “registration dictionary” to display on the display function unit that the character string unit is a new character string and to store the result in the analysis result storage function unit. Here, the term “new character string” means that the character string is not stored (registered) in the proofreading dictionary or the correction history registration dictionary. If the character string is a new character string, the display function unit displays it. After this display is finished, the analysis position is moved to the end position of the new character string and the analysis is resumed.
A text update function unit 6 replaces the proofreading expression with an alternative candidate, and updates the text stored in the text storage function unit 1 according to a new input. This determination can be made automatically or according to the instructions of the proofreader.
Reference numeral 7 denotes a mode designating unit which activates the text update function unit 6 in response to an instruction from the text update instruction function unit 8 when the text is updated. As soon as an interactive mode or an alternative expression for replacement is found, the sentence update function unit 6 is activated and the proofreader selects whether to set the automatic mode for automatic replacement.
Reference numeral 8 denotes a text update instruction function unit, which is a function unit that instructs the proofreader to update the text.
A control function unit 9 is a function unit that activates the text update function unit 6 and the text update function unit 8 according to the mode set by the mode designating unit 7.
An alternative expression input function unit 10 is a function unit that receives an alternative expression corresponding to a proofreading target expression input by a proofreader.
A registration instruction function unit 11 is a function unit that allows the proofreader to select whether or not to register the proofreading target expression (character string) and its alternative candidate (character string) in the correction history registration dictionary 3. . When the proofreader selects manual correction or automatic correction and sends the instruction to the registration instruction function unit, when manual correction is instructed, a display waiting for the instruction from the proofreader is displayed. When the automatic correction is instructed as shown in the display function unit, the correction is automatically performed by the system without the manual of the proofreader.
Reference numeral 12 denotes a correction history registration function unit (dictionary), which is a function unit for registering the proofreading target expression calibrated and the alternative candidate in association with each other in the correction history registration dictionary 3. Actually, in addition to these function units, in addition to the text input function unit (not shown) for the proofreader to input Japanese text, it also displays Japanese text and detected proofreading expression etc. Although various components such as a display function unit are required, they are not the main point of the present invention, and thus the description thereof is omitted.
The following description will be given by taking as an example a case where an error in a sending kana as a proofreading expression is calibrated with respect to a sentence input in advance for the sentence proofreading apparatus configured as described above. It should be noted that other proofreading expressions such as inconsistencies in sending kana and prohibited terms can be proofread by a similar operation.
(1) At the time of sentence proofreading, it is set by the proofreader whether to proofread the sentence interactively or automatically through the mode designating unit 7.
Here, this mode is set with the intention that the proofreader automatically proofreads.
(2) In order to proofread a sentence that has been input in advance, the proofreader gives an instruction to detect an error in the kana.
(3) The proofreading target expression detection function unit 2 detects an error in the sending kana from the Japanese sentences stored in the sentence storage function unit 1. As a result of the detection, it is assumed that the expression to be calibrated “Let's go” is detected.
(4) The substitution candidate determination function unit 5 first activates the correction history search function unit 4 in order to determine a substitution candidate for correcting an error in the sending kana.
(5) The correction history search function unit 4 searches the correction history registration dictionary 3 using the proofreading expression “Yoko Konau” as a search key, and has a history that the expression “Yoko Konau” has been previously corrected to another expression. Check if there is any. Here, information indicating that the corresponding item could not be searched is output as a case where the corresponding correction has not been performed before.
(6) The alternative candidate determination function unit 5 activates the alternative expression input function unit 10 to accept the input of the alternative expression in accordance with the information from the correction history search function unit 4 described above.
(7) The alternative expression input function unit 10 prompts the proofreader to input an alternative expression for the expression to be proofread, and outputs the input alternative expression. For example, if “perform” is input as an alternative expression, the registration instruction function unit 11 is activated to receive an instruction from the proofreader as to whether or not the correction history should be registered.
(8) The registration instruction function unit 11 inquires of the proofreader whether or not the correction history should be stored in the correction history registration dictionary 3, and receives an instruction from the proofreader and outputs it. At this time, if the proofreader determines that the correction is temporary, for example, the proofreader decides not to register, and the correction is not repeated but is detected several times thereafter and the same correction is performed. If it is determined that there is a possibility, “registration” may be determined. When an instruction “register” is given from the proofreader, the registration instruction function unit 11 accepts the instruction and outputs it to the correction history registration function unit 12.
As an aid to determining whether the proofreader is “temporary” above, if the system detects that there should be multiple similar “errors” on the same screen, The system is preferably organized so that they are all displayed simultaneously. This multiple display allows the proofreader to predict that this “error” will appear in the document at least in the future. In addition, if the character string to be corrected is a word to be used, the ending change of each word may not always be the same depending on its position, so they are optimally adjusted to the context For this reason, it is useful that each “error” is displayed on the same screen.
(9) The correction history registration function unit 12 makes a correction by associating the proofreading target expression character string “Yokona” with the alternative candidate character string “Done” for the correction in accordance with the output from the registration instruction function unit 11. Register in the history registration dictionary 3. At this time, if the same thing to be registered is already registered in the correction history registration dictionary 3, registration is not performed.
(10) The proofreader uses the sentence update instruction function unit 8 to instruct that to correct the expression to be proofread in the sentence.
(11) The text update function unit 6 updates the content of the text storage function unit 1 by rewriting the proofreading target expression “Gone” to the alternative candidate “perform” input by the alternative expression input function unit 10.
By each of the above processes (4) to (11), the correction operation for the proofreading target expression “Konako” is completed. Further, the detection and correction work of the sending kana error is continued for the succeeding sentence. Then, as described above, the same proofreading expression “Yokona” is detected again in a state where the proofreading expression “Yokokono” and the alternative expression “Perform” are registered in the correction history registration dictionary 3. In such a case, the following operation is performed.
(12) The alternative candidate determination function unit 5 first activates the correction history search function unit 4 in order to determine an alternative candidate for correcting the proofreading target expression “Konaru”.
(13) The correction history search function unit 4 searches the correction history registration dictionary 3 by using “Let's go” as a search key. Here, since the correction history registration function unit previously registered “correction history” using “do” for the expression to be calibrated “going”, “do” is searched and the information is output. The
Therefore, when the automatic mode is set by the mode designating unit 7 described above, the functions or operations shown in the following items are executed.
(14) The alternative candidate determination function unit 5 activates the text update function unit 6 to perform the same correction as the past correction from the output of the correction history search function unit 4.
(15) The text update function unit 6 updates the content of the text storage function unit 1 by rewriting the proofreading target expression to an alternative candidate. If the update is performed automatically at that time, no instruction from the proofreader is required.
(16) The document update function unit 6 is activated by an instruction from the proofreader from the document update instruction function unit, and is replaced with alternative candidates while sequentially checking the proofreading target expression.
In the case of a contribution that can be automatically proofread as described above, and when making a correction that is exactly the same as a correction that has already been made in the past, the proofreader inputs an alternative candidate and instructs the update of the text. The operation can be omitted, the operation can be simplified and the correction time can be shortened.
In addition, it is possible to apply the same correction to the same proofreading expression, and in the correction work stage, a new error (inconsistent notation, expression with another word of the same concept, etc.) will be committed. Can be eliminated. Further, by providing the registration instruction function unit 11, it is possible to selectively store only the defense history judged to be necessary by the proofreader, and it is possible to specify a correction candidate only after judging including the context before and after. In this case, it is possible to avoid making an unintentional correction automatically.
In addition to this, in the present invention, by providing the mode designating unit 7, in addition to the method of automatically correcting the proofreading object expression, the proofreader expression that has been performed in the past is corrected while sequentially checking. Both manual methods can be adopted. In this embodiment, an example in which a character string of an alternative candidate is used as information representing an alternative candidate is shown, but other information such as a series of instructions for correcting index information in the proofreading target dictionary or the character string is shown. Information can also be used.
In addition, although an example using the example of the proofreading target expression as information representing the proofreading target expression is shown, “other information such as index information of the proofreading target dictionary or position and length information in the text storage unit 21 is shown. Can also be used.

(Example 2)
Even if there is “... dealed ...” in the text to be proofread, this error cannot be corrected automatically by the same document proofreading system used in the first embodiment. That is, the system cannot suggest how to correct this error. This is because there is nothing outside the language knowledge stored in the brains of proofreaders who use human word processing software. Although this error was supposed to be “used” in the regular expression, it is presumed that the error was caused by a malfunction that occurred in the finger usage of the word processor proofreader.
Since it is almost impossible to predict this type of error, it is not possible to register such a typo in the proofreading dictionary in advance. Predicting whether or not this error will reappear at the time of encountering such a mistake, and if it can be predicted that there is a high probability of reappearance, this typographical phrase (misrepresentation) will be stored in the proofreading dictionary. Once registered, the document proofing system will automatically detect and display the next occurrence. If the mode specifying unit 7 is set to the automatic mode, the document proofreading system can automatically execute the steps (14) to (16) described above to automatically correct the error. .
In the present invention, as a specific method example for creating an improved proofreading dictionary, a combination of a searched character string and an arbitrary searched character can be adopted. In the word processor software "Ichitaro", "\?" For a line break, the “Ichitaro” macro uses “\ n” as a line feed (enter) mark in the searched character string for “Ichitaro”. In practice, the search string defined in this SearchString is registered in the form of Alt + Roman letter, for example "V". The number of characters that can be registered is approximately 50 according to the CPU's ability. Therefore, when registering search character strings exceeding this number, change the Roman character combined with Alt to another character, for example, “C”. Registration will continue.
In the second embodiment, Alt + V is registered in the system as a registration key, and for example, SearchString (.search string = "different \ | is unexpectedly expressed as \ | \ | Painting \ | etc. can be \ | can be \ | or can be \\ or \\ information \ | public information \ | recycle ",. Search direction = 1 ,. fuzzy search = 0) was registered. “\ |” Corresponds to “or” search.
When an actual sentence is proofread using this SearchString, a lexical phrase (a character string pre-registered by each proofreader) preliminarily registered by each proofreader listed above is always detected. As a result, the detected character string can be surely corrected by the proofreading system automatically or according to instructions from the proofreader.

As described above, the present invention can achieve the following effects:
(1) Correction history search function unit and proofreading that searches for an alternative expression based on a correction history registration dictionary in which the proofreading expression and the alternative expression replaced with the proofreading expression are registered at the time of proofreading By providing a replacement function unit that automatically replaces the target expression with the alternative candidate, the sentence can be automatically proofread, and when the same correction as previously made is performed The proofreader can omit both the operation of selecting an alternative candidate and the operation of instructing the update of the text. With this contribution, in addition to being able to proofread a sentence with a simple operation, it is also possible to produce an effect of shortening the correction time at the same time.
(2) In other words, this correction operation unifies narrower synonyms (usually multiple synonyms that are only different from each other) in the same document even if they appear multiple times. Useful for purposes. For example, “appeal” and “complaint”, “procedure” and “procedure”, etc., need to be unified to a predetermined character before the documents are submitted. These points are not only due to inconsistencies, but also because they are required to be terms defined in laws, regulations or regulations. In addition, it should be avoided that “but” and “but” coexist in the sentence unless there is a special meaning. However, “time” and “time” should not be unified in legal documents. This is because they have different legal meanings. In general, law books, judgment documents, provisional disposition orders, execution orders, etc., are often given different meanings from hiragana writing and kanji writing. You should avoid doing it. As another type of problem, it is wise to avoid unifying the use of letters in the old law and the new law. In the law enacted before the war, script terms were selected without being restricted by kanji, so the new law enacted under the restriction by the kanji after the war may have changed to different script terms. In particular, it should be noted that, in a sentence quoting the provisions of the old law, it may lead to an unexpected error to match the old law script terms to those of the new law.

The block diagram of the text proofreading apparatus in the Example of this invention. The schematic diagram which shows one example of the content of the correction log | history registration dictionary in the Example. The block diagram of the conventional text proofreading apparatus.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Text memory function part 2 Calibration object expression detection function part 3 Correction history registration dictionary 4 Correction history search function part 5 Alternative candidate determination function part 6 Text update function part 7 Mode designation part 8 Text update instruction function part 9 Control function part 10 Alternative Expression input function unit 11 Registration instruction function unit 12 Correction history registration function unit 21 Text storage unit 22 Calibration target expression detection unit 23 Alternative expression dictionary memory 24 Alternative expression display control unit 25 Correction correction control unit 26 Input unit 27 Display unit

Claims (13)

  A document proofreading system that detects one or more character strings registered in advance by a proofreader and recognized by the proofreader as requiring correction from the entire proofreading sentence or a designated part. A sentence proofreading method characterized by displaying a detected character string and instructing a proofreader to correct the character string as necessary after confirmation.   The proofreader selects a substitute expression from the displayed alternative candidates and corrects the displayed character string when the proofreader instructs correction of the displayed character string. The written proofreading method.   The proofreader inputs an alternative expression and corrects the displayed character string when an alternative candidate is not displayed when the proofreader instructs correction of the displayed character string. Item 3. The sentence proofreading method according to item 1 or 2.   3. The displayed character string is corrected by automatically replacing the character string with an alternative expression when the displayed character string meets a predetermined correction condition. The written proofreading method. An input function unit, a proofreading target expression detection function unit that detects a proofreading target expression in the proofreading text, and
A correction history registration dictionary in which both the proofreading target expression and the proofreading target expression and the alternative expression replaced at the time of sentence proofreading are registered in association with each other;
A correction history search function unit, which is basically composed of a function unit and a display function unit for searching for an alternative expression using the correction target expression detected by the correction target expression detection function unit based on the correction history registration dictionary as a search key. Sentence proofreading device characterized by being organized into.   The sentence proofreading apparatus further includes a function unit for determining an alternative candidate from an alternative expression registered in the correction history registration dictionary in association with the proofreading expression in the correction history registration dictionary. Item 6. The sentence proofreading apparatus according to Item 5.   The sentence proofreading apparatus further causes the proofreading target character string to be corrected using the alternative expression input by the proofreader as an alternative candidate when the alternative expression is not registered in the correction history registration dictionary. Sentence proofing device.   The sentence proofreading apparatus further includes a function unit that automatically replaces and corrects the proofreading target expression and the alternative candidate determined by the alternative candidate determination function unit as a replacement function unit. Item 6. The sentence proofreading apparatus according to Item 5.   The sentence proofreading apparatus is further provided with a function unit that is a correction history registration function unit and associates the replaced proofreading target expression with an alternative candidate and additionally registers it in the correction history registration dictionary. The sentence proofreading apparatus in any one of 5-8.   The sentence proofreading apparatus according to any one of claims 5 to 9, further comprising a selection function unit that selects whether or not to register with the correction history registration function unit in accordance with an instruction from a proofreader.   The sentence proofreading apparatus according to claim 5, further comprising a selection function unit that selects whether or not to register with the correction history registration function unit in accordance with an instruction from a proofreader.   12. The sentence according to claim 5, further comprising a mode switching function part which is automatically performed when the replacement function part is instructed by a proofreader. Calibration device.   13. The sentence proofreading apparatus according to claim 12, wherein the mode switching function section includes a function section for switching the mode so that the mode switching function section is automatically activated when the replacement function section is instructed by a proofreader.