JPH07271791A - Document editing device - Google Patents

Abstract

PURPOSE:To correct only a word irrelevantly to the number of its characters by recognizing and segmenting the word before correction by a syntax analysis at the time of the correction. CONSTITUTION:This document editing device is equipped with an editing part 26 including a cursor painter indicating a cursor position in a document and a correction insertion mode flag storing the mode of correction insertion, a syntax analysis part 27 including a syntax analytic dictionary for analyzing and decomposing a KANJI (Chinese character) - KANA (Japanese syllabary) mixed sentence, an analytic buffer for storing the analytic result, a classification buffer for storing its classification, a part-of-speech buffer storing its part of speech, and an inflection type buffer for storing its inflection type, a KANA- KANJI conversion part 28 including a KANA-KANJI conversion dictionary for converting inputted KANA into KANJI, a a conversion buffer for storing the conversion result, a learning buffer for storing its learning, a classification buffer for storing its classification, a part-of-speech buffer for storing its part of speech, and an inflection type buffer for storing its inflection type, and a flag 29 which indicates whether or not automatic correction is performed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a Japanese word processor, DTP (Desk Top Publish).
g) and other grammar correction devices.

[0002]

2. Description of the Related Art Correction (overwriting) of a conventional Japanese word processor is performed by a user as follows when changing "execute" to "execute" or "execute" to "execute". It is processed in steps.

FIG. 13 is an overall configuration diagram of a conventional technique.
Reference numeral 11 is a CPU, 12 is an input unit such as a keyboard for inputting characters and the like, 13 is a printing unit, 14 is an external storage unit such as FD, 15 is a display unit such as CRT, 16 is a document buffer for storing the input sentence, a document A cursor pointer indicating the cursor position, an editing unit including a correction insertion mode flag that stores a correction insertion mode, 17 is a Kana-Kanji conversion dictionary for converting an input Kana into Kanji, a conversion buffer that stores the converted result, It is a kana-kanji conversion unit that includes a learning buffer that stores the learning.

FIG. 3A shows an example of insertion. Insertion is an editing process in which a character is added at the position where the cursor is. (2) is an example of correction. The correction is an editing process for replacing the current character and the input character from the cursor position by matching the number of characters. Deleting one character means deleting one character at the cursor position. When erased, the next character will automatically advance to the current cursor position.

FIG. 14 shows a conventional operation screen for correction.
In the case of “execute” to “execute”, the cursor is moved to the “line” with the keyboard 12, “execute” is input in the insert mode, and “execute” is erased by deleting one character. Or
When you move the cursor over "Line" and enter in Hiragana, "Okay" (4 characters) and "Perform" (6 characters)
Enter two spaces to match the number of characters
Move the cursor over the space of the character space, switch to the correction (overwrite) mode, input "Jiko", and convert it into "Execute" by the Kana-Kanji conversion unit 17. In this case, if the numbers of characters are not matched, the 6 characters following "Perform" will be corrected.

In the case of "execute" to "execute", move the cursor over "execute", enter "execute" in the insert mode, and erase "execute" by deleting one character. Alternatively, move the cursor over "Actual", switch to the correction (overwrite) mode, enter "Perform", and erase the remaining "Perform" by correction (overwrite) by erasing one character.

As described above, in the correction (overwriting), the user compares the characters before and after the correction in his head, and performs the work of erasing and inserting.

Further, in the past, Japanese Patent Laid-Open No. 63-289667
(Name "Japanese document editing device") discloses that the control unit is provided with a character type discriminating unit, and batch designation of a character string is repeatedly performed with a position where the character type changes as a boundary. I judge, but in this case I can not judge correctly in the case of verbs and adjectives. For example,
In the case of “do”, “line” and “u” are treated as different characters because they are different in kanji and hiragana character types.

As another conventional example, Japanese Patent Laid-Open No. 4-528
54 (name “document data processing device”) discloses that a corrected character string and a corrected character string are automatically replaced with each other. In this case, for example, “our company” is “in our company”. If there is no same "," at the end like ",", it cannot be processed. "Do" cannot be corrected as "execute."

FIG. 15 shows an example of a general dictionary. The higher the number, the higher the frequency. The connected dictionary is
The connected words in the homonyms are output as candidates with priority over the frequency of the homonyms by the preceding and succeeding words.

In kana-kanji conversion of character input, the order of priority of homonyms is to add a general frequency when creating a dictionary and to issue homonyms in that order ("high school""takayuki""navigation"). )) And the method of putting out what you learned once with the highest priority next time (If you select "Navigation", next time "Navigation""Highschool"
The method of checking the connection between words and giving priority to the connected words ("baseball" as the next word of "high school" and "filament" as the next word of "parent") Are being considered.

[0012]

As described above, in the conventional example, it is necessary for the user to be aware of the character lengths before and after replacement and correct (overwrite). In kana-kanji conversion at the time of correction, learning was the only effective method for selecting homophones, and there was no other method.

The present invention has been made in view of such circumstances, and corrects (overwrites) "execute" to "execute".
When doing, the machine recognizes the numbers 2 and 4 of “do” and “execute”, and the sentence “when doing” is automatically 2 to 4 characters “when doing” without the user being aware of it. The purpose is to correct (overwrite). Furthermore, when correcting (overwriting) “scan” to “scan”, the usual kana-kanji conversion is “manipulate” and “investigate”.
It is necessary to press the next candidate several times to select the desired word such as "scan", but if "scan" and "scan" are in the same category, "scan" is the first in kana-kanji conversion. 1
It is intended to be converted as a candidate. Therefore, the input is very easy.

[0014]

According to the invention described in claim 1, input means for inputting characters, document storing means for storing the input characters, display means for displaying the input characters, and input Cursor display means for indicating the position to be moved, syntactic analysis means for decomposing and analyzing kana-kanji mixed sentences in word units,
Equipped with a kana-kanji conversion means for converting the input kana into kana-kanji, and if the cursor is on the already entered character and the input mode is the correction (overwrite) mode, one sentence at the cursor position is parsed. It is a text editing device characterized by confirming the part of speech of a word on which a cursor is located and adjusting the number of characters replaced by the word in the syntax analysis by means.

In this case, it is also possible to confirm the inflectional form of the word on which the cursor is located and automatically match the inflectional form of the word to be replaced.

According to the second aspect of the invention, input means for inputting characters, document storage means for storing the input characters, display means for displaying the input characters, and cursor display for indicating the input position Means, a syntactic analysis means for decomposing and analyzing kana-kanji mixed sentences in word units, and a kana-kanji conversion means for converting an inputted kana into kana-kanji characters, and there is a cursor on an already-input character. When the mode is the correction (overwrite) mode, one sentence at the cursor position is parsed by the parsing means, the classification of the word at the cursor is confirmed, and the kana-kanji conversion means is used to check the relationship between the original word and the synonyms. The sentence editing apparatus is characterized in that the word in the is a first candidate.

In this case, it is also possible to confirm the classification of the word on which the cursor is located and use the word having an antonym relation with the original word as the first candidate by kana-kanji conversion.

Furthermore, it is possible to confirm the classification of the word with the cursor and switch the learning priority order between the original word and the synonyms, and the antonym related words by kana-kanji conversion.

Further, it is possible to simplify the speed and the analyzing means by analyzing the word with the character at the cursor position as the head instead of the whole sentence.

According to the third aspect of the invention, input means for inputting characters, document storing means for storing the input characters, display means for displaying the input characters, and cursor display for indicating the input position Means, a syntactic analysis means for decomposing and analyzing kana-kanji mixed sentences in word units, and a kana-kanji conversion means for converting an inputted kana into kana-kanji characters, and there is a cursor on an already-input character. When the mode is the correction (overwrite) mode, one sentence at the cursor position is syntactically analyzed by the syntactic analysis means, and in the case of input by the abbreviated conversion, when the abbreviated conversion is registered in the dictionary part, the part of speech at the time of kana-kanji conversion The text editing device is characterized in that the same correction as in the kana-kanji conversion can be made in the abbreviated conversion by adding the information classification information.

Further, in the inventions according to claims 1, 2, and 3, it is possible to provide a text editing device in which a user can switch whether or not to perform these processes.

[0022]

According to the invention described in claim 1, the machine (CPU 21) detects the part of speech or classification of the word on which the cursor is located and recognizes the difference in the number of characters, thereby confirming the part of speech of the word on which the cursor is located. Then, the number of characters replaced with the word is adjusted. Further, it becomes possible to confirm the inflectional form of the word at the cursor and automatically match the inflectional form of the word to be replaced.

According to the second aspect of the present invention, the classification of the word on which the cursor is located is confirmed, and the word having a relationship between the original word and the synonym in kana-kanji conversion is set as the first candidate. Furthermore, the classification of the word with the cursor is confirmed, and the word having an antonym relation with the original word by kana-kanji conversion is set as the first candidate. Alternatively, it is possible to confirm the classification of the word where the cursor is, and switch the learning priority between the original word and the synonyms, and the antonym related words by kana-kanji conversion.

In this case, it is also possible to simplify the speed and the analyzing means by analyzing the word with the character at the cursor position as the head instead of the whole sentence.

According to the invention described in claim 3, when the short conversion is used as the input method, by adding the part-of-speech information classification information at the time of the kana-kanji conversion at the time of registering the short conversion in the dictionary part, The same correction as in the kana-kanji conversion can be made even in the abbreviated conversion.

Alternatively, it is possible for the user to switch whether or not to perform the above processing.

[0027]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below in detail with reference to the embodiments shown in the drawings. The present invention is not limited to this.

FIG. 1 is a block diagram showing the configuration of a text editing apparatus according to an embodiment of the present invention. Reference numeral 21 is a CPU, 22 is an input unit such as a keyboard for inputting characters, 23 is a printing unit, 24 is an external storage unit such as FD, 25 is a display unit such as CRT, 26 is a document buffer for storing the input text,
A cursor pointer indicating the cursor position of the document, an editing unit including a correction insertion mode flag for storing the correction insertion mode, 27 is a syntactic analysis dictionary for analyzing and decomposing a kanji kana mixed sentence, an analysis buffer for storing the analysis result, and A classification buffer that stores a classification, a part-of-speech buffer that stores the part-of-speech, a parsing unit that includes a usage-type buffer that stores the inflection, 28 is a kana-kanji conversion dictionary for converting an input kana into kanji, and the conversion result is stored. A conversion buffer, a learning buffer that stores the learning, a classification buffer that stores the classification, a part-of-speech buffer that stores the part-of-speech, a kana-kanji conversion unit that includes the inflectional-type buffer that stores the inflection, and 29 is a flag that indicates whether to perform automatic correction. Is.

FIG. 2 shows an example of correction. Usually the parts of speech are the same, but the number of characters is often different. First, "patent" is replaced with "utility model", but the particle "no" is common, so the noun "patent" is replaced with the noun "utility model".

FIG. 3 shows an example of the correction screen. (1) is an insert mode for adding sentences, (2) is a conventional correction example,
(3) is a correction example of the present invention. In the second correction, if the number of characters is different, the number of characters that are not desired to be corrected (in the example, "ha") will also be corrected, so that they will be corrected. In the present invention, post-processing is unnecessary and efficient input is possible.

4 and 5 are dictionaries for kana-kanji conversion and syntax analysis. The part of speech of the dictionary includes nouns, verbs, adjectives, adjectives, adverbs, conjunctions, etc.

The kana-kanji conversion dictionary extracts kana, part-of-speech, and classification from the kana and analyzes them. A grammatically correct combination of kanji and kana sentences is displayed as a result.

The syntactic analysis dictionary reads from Chinese characters, extracts a part of speech and a classification, and analyzes them. Display grammatically incorrect combinations, unparseable words, etc.

The inflectional dictionary includes verbs, adjectives, adjective verbs,
Has the use of auxiliary verbs. It is processed as a single word by applying a conjugation to the stem of the main dictionary.

In the case of "Aiba", the headline "A" (use of the part-of-speech verb Wawa 5 lines) of the Kana-Kanji conversion dictionary, and the hypothesis "E" and the particle "Ba" of the use form 5 lines Is judged to have been connected.

FIG. 6 shows an example of a combination of parts of speech at the time of correction. The same part-of-speech may be used in the same usage. To correct a verb into a noun, use the plural form or add a "koto" to make it a noun. To use an adjective as a noun, use "sa".
To make a noun.

FIG. 7 shows an example of syntax analysis. First, a kanji-kana mixed sentence is decomposed into words such as nouns, verbs, and particles, and nouns and particles, verbs and auxiliary verbs, etc. are collected and combined as clauses.

FIG. 8 shows an example of kana-kanji conversion. The input kana is decomposed into words such as nouns, particles, and verbs, and the nouns and verbs are converted into kana-kanji characters.If there are homophones, the learning buffer is searched, and they are replaced in the candidate order. The original classification is investigated, and if there is a word of the same classification, that word is set as the first candidate for the homophone. Then, a noun and a particle, and a verb and an auxiliary verb are combined to form a clause.

FIG. 9 is an overall flow chart. First, the mode is checked, and if it is the insertion mode, the process proceeds as it is. If it is the correction mode, it is checked whether it is the automatic correction mode. In the automatic correction mode, the original sentence is parsed and the part of speech and classification of the word at which the cursor is located are extracted ("do" is analyzed as a verb). Characters are entered ("Perform"
When the Kana-Kanji conversion key is pressed (it becomes "execute"), the part of speech of the conversion result is taken out (sa verb) and compared with the part of speech of the correction source word. Check, adjust the inflectional form (both are the final form or the adnomial form, in this case there is a noun behind, so the adjunct form), delete the correction source word (delete "do"), and insert the correction word (Insert "execute").

Correct "overwrite" to "execute"
When doing, the difference between the numbers of characters 2 and 4 is detected by the machine (CPU) by detecting the parts of speech and classification of the words “do” and “execute”.
21) recognizes, and the sentence "when performing" is automatically changed from 2 letters to 4 when "when performing" without the user being aware of it.
It corrects (overwrites) the characters.

FIG. 10 is a flow chart of parsing. First, clear the previous word information, search for the word with the longest match (search for "line" (noun) and "line" (verb wah line, but consider both with the longest match) and find the word For example, check if it is a idiom (“do” is a idiom), and if it is a idiom, search for the following characters as the ending (“u” is the verb w
Judged as a union form). Next, the particle is searched (there is no particle in this case), and the connection with the previous word is checked. If the connection is OK, it is stored in the analysis buffer. This is repeated until one sentence is completed. If not connected, the word length is shortened by one character and continued.

"To do" is decomposed into the adnominal form of the verb "to do" and the noun "to". Since the noun "row" and "u" are not connected by the connection check, they are deleted from the candidates.

FIG. 11 is a flowchart of kana-kanji conversion. When a character is input (“Yes” is input), it is first checked whether it is the next candidate, and if it is the next candidate, the next homophone is taken out from the homophone buffer. In the case of Kana-Kanji conversion, the word is searched from the Kana-Kanji conversion dictionary with the longest match (the noun "so", the noun "so", the "so"
Of the nouns, and the verb of the "sousa" and the noun of the syllabic words of "sousa" are selected). Next, it is checked whether it is a phrasal (sa verb "Sosa" is a verb), and if it is a verb, the ending of the word is searched ("shi" is a continuation form of the verb), and then a particle is searched. If the connection is OK, the result is stored in the result buffer (“sousashi” is stored as a continuous form of the verb). If there are homophones, the homonyms are sorted by learning (generally in the order of “manipulate”, “investigate”, and “scan”). If the correction is in automatic mode, take out the classification of the original word (“scan” is a continuous form of the verb, and the classification is in the field of electrical engineering),
The order of homonyms of the same classification as the original word is the first candidate ("operation" and "investigation" are general fields, and the field of "scanning" is electrical engineering, so the order of "scanning", "operation", and "investigation" is Replace with). In the case of selecting a homophone, write learning.

When "scan" is corrected (overwritten) to "scan", the normal kana-kanji conversion is "operate".
It is necessary to press the next candidate several times to select a desired word such as “investigate” and “scan”, but when “scan” and “scan” are in the same classification as in the present invention, Kana-Kanji conversion converts "scan" as the first candidate, which makes input much easier.

When the character at the cursor position is analyzed as the beginning of a word, it is the same as the word conversion by kana-kanji conversion and the consecutive phrase conversion, and it is not necessary to analyze from the beginning of one sentence.
Analysis time is short. Moreover, since a dictionary for parsing is unnecessary, it can be realized at low cost.

If there is a simple reverse lookup dictionary corresponding to the word conversion dictionary for Kana-Kanji conversion, the part of speech and field of that word can be extracted.

FIG. 12 shows an example of a dictionary structure for shortening conversion.
In the abbreviation conversion, the characters that have been input once are stored in pairs as the first kana and kana-kanji converted words, and if you want to input the same word next time, enter only the first kana character and convert it. It's a convenient thing that you can enter just by pressing. Previously, I remembered only one kana character and word (stem),
In the present application, by storing the part-of-speech information and the classification information obtained by the first Kana-Kanji conversion, the shortening conversion dictionary has the same format as a normal Kana-Kanji conversion dictionary even when the shortening conversion is used. You can check the ending change from the part-of-speech information as in the case of Kana-to-Kanji conversion.

Further, a dictionary which can be freely added by the user A kana-kanji conversion dictionary which is generally called a user dictionary does not differ from a normal kana-kanji conversion dictionary if the part-of-speech information is properly entered. You can check the ending change from the part-of-speech information as in the case of normal Kana-Kanji conversion.

[0049]

According to the invention as set forth in claim 1, the machine (CPU 21) detects the part of speech, classification, etc. of the word on which the cursor is located and recognizes the difference in the number of characters, so that the part of speech of the word on which the cursor is located. And adjust the number of characters replaced by the word. Further, the inflectional form of the word at the cursor is confirmed, and the inflectional form of the word to be replaced is automatically adjusted. Thus, at the time of correction, there is an effect that the word before correction is recognized by the syntax analysis, cut out, and only that word is corrected regardless of the number of characters.

According to the second aspect of the present invention, the classification of the word on which the cursor is located is confirmed, and the word having a relationship between the original word and the synonym in the kana-kanji conversion is set as the first candidate. In addition, in the kana-kanji conversion when inputting the corrected characters, in order to give priority to those that have a synonym or antonym relationship between the word before correction and the meaning, check the classification of the word with the cursor A word having an antonym relationship with the original word in the conversion is set as the first candidate. Or check the classification of the word the cursor is on,
The kana-kanji conversion can switch the learning priority between the original word and the synonyms, and the antonym related words.

In this case, further, it is possible to simplify the speed and analysis means by analyzing the word with the character at the cursor position as the head instead of the whole sentence.

According to the invention as set forth in claim 3, in the case of input by the abbreviated conversion, by adding the part-of-speech information classification information at the time of the kana kanji conversion at the time of registration in the dictionary part of the abbreviated conversion, the abbreviated kana can also be obtained. It enables the same correction as Kanji conversion.

Alternatively, there is an effect that the user can switch whether or not to perform the above processing.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a text editing device according to an embodiment of the present invention.

FIG. 2 is a diagram showing a correction example in the text editing apparatus according to the embodiment of the present invention.

FIG. 3 is a diagram showing an example of a correction screen in the text editing device according to the embodiment of the present invention.

FIG. 4 is a drawing showing an example of a dictionary in a text editing device according to an embodiment of the present invention.

FIG. 5 is a drawing showing an example of a dictionary in the text editing device of one embodiment of the present invention.

FIG. 6 is a drawing showing an example of a combination of parts of speech at the time of correction in the text editing device of one embodiment of the present invention.

FIG. 7 is a drawing showing an example of syntax analysis in the text editing device of one embodiment of the present invention.

FIG. 8 is a diagram showing an example of kana-kanji conversion in the text editing device according to the embodiment of the present invention.

FIG. 9 is a flowchart showing the overall operation of the text editing device of one embodiment of the present invention.

FIG. 10 is a flowchart showing an operation of syntax analysis in the text editing device according to the exemplary embodiment of the present invention.

FIG. 11 is a flowchart showing an operation of kana-kanji conversion in the text editing device according to the embodiment of the present invention.

FIG. 12 is a drawing showing a shortening conversion dictionary in a text editing device according to an embodiment of the present invention.

FIG. 13 is a block diagram showing a configuration of a conventional text editing device.

FIG. 14 is a diagram showing an example of a correction screen in a conventional text editing device.

FIG. 15 is a drawing showing an example of a dictionary in a conventional text editing device.

[Explanation of symbols]

 21 CPU 22 Input Section 23 Printing Section 24 External Storage Section 25 Display Section 26 Editing Section 27 Syntax Analysis Section 28 Kana-Kanji Conversion Section 29 Automatic Correction Mode Flag

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI technical display location 9288-5L 526A

Claims (3)

[Claims] 1. An input unit for inputting characters, a document storage unit for storing the input characters, a display unit for displaying the input characters, a cursor display unit for indicating an input position, and a kana-kanji mixed sentence. It is equipped with a syntax analysis means for decomposing and analyzing word by word and a kana-kanji conversion means for converting the input kana into kana-kanji characters. There is a cursor on a character that has already been input, and the input mode is the correction (overwrite) mode. In this case, a sentence editing device, wherein one sentence at the cursor position is syntactically analyzed by the syntactic analysis means, the part of speech of the word at the cursor is confirmed, and the number of characters replaced by the word is adjusted. 2. An input means for inputting characters, a document storage means for storing the input characters, a display means for displaying the input characters, a cursor display means for indicating an input position, and a kana-kanji mixed sentence. It is equipped with a syntax analysis means for decomposing and analyzing word by word and a kana-kanji conversion means for converting the input kana into kana-kanji characters. There is a cursor on a character that has already been input, and the input mode is the correction (overwrite) mode. In this case, in the case where one sentence at the cursor position is syntactically analyzed by the syntactic analysis means, the classification of the word with the cursor is confirmed, and the kana-kanji conversion means sets the word having the relationship between the original word and the synonyms as the first candidate. A text editing device characterized by: 3. An input unit for inputting characters, a document storage unit for storing the input characters, a display unit for displaying the input characters, a cursor display unit for indicating an input position, and a kana-kanji mixed sentence. It is equipped with a syntax analysis means for decomposing and analyzing word by word and a kana-kanji conversion means for converting the input kana into kana-kanji characters. There is a cursor on a character that has already been input, and the input mode is the correction (overwrite) mode. In this case, in the case of parsing one sentence at the cursor position by the parsing means, by adding the part-of-speech information classification information at the time of kana-kanji conversion at the time of registration in the dictionary part of the abbreviated conversion when inputting by the abbreviated conversion , A text editing device characterized in that the same correction as in kana-kanji conversion can be made even in abbreviated conversion.