JPH11134337A - Notation indicating device and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To appropriately assist document proofreading, to reduce the load of word registration of KATAKANA (square form of Japanese syllabary) words, and to suppress excessive indication by providing a means which takes out a wrong morpheme registered in a dictionary for analysis and a means which classifies morphemes by error levels and makes an emphasized display. SOLUTION: A morpheme analyzing means 2 analyzes and divides an inputted document into morphemes and takes out wrong morphemes (proofread word) registered in the dictionary 6 for morpheme analysis, and a detection result output means 5 classifies errors by levels as to the taken-out proofread words and makes their emphasized display, and also displays replacing candidates. Further, a contiguous character kind deciding and identical character kind acquiring means 4 puts together contiguous word having the same attribute in the document as to the taken-out proofread words and the detection result output means 5 puts together words after composition and classifies errors by levels, and makes their emphasized display and displays replacing candidates. Consequently, the load of word registration of KATAKANA words, etc., is lightened and excessive indication can be suppressed.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a notation indicating device for indicating an error in notation of a character string in a sentence and a recording medium.

[0002]

2. Description of the Related Art Conventionally, when detecting and pointing out a spelling error by performing Japanese processing, it is necessary to construct a dictionary for detecting the spelling error. However, since it is virtually impossible to register all the correct notations in the dictionary, they pointed out as unregistered words regardless of the notations written correctly.

[0003]

As described above, when an error in a conventional notation is detected and pointed out, it is necessary to register all correct notations in a dictionary, and it is difficult to create a dictionary. There were problems such as being pointed out as unregistered words irrespective of the notation correctly written because it could not be registered in the dictionary.

[0004] The present invention solves these problems,
After preparing an analysis dictionary and extracting the morpheme (correction word) of the error from the input sentence and synthesizing adjacent words with the same attribute, dividing the level and highlighting indications and replacement candidates highlighted according to the error level of the word The purpose is to display and provide appropriate support for grammar correction, to reduce the burden of registering words such as katakana, and to suppress excessive pointing.

[0005]

Means for solving the problem will be described with reference to FIG. In FIG. 1, a morphological analysis unit 2 morphologically analyzes an input sentence to divide it into morphemes, and refers to a morphological analysis dictionary 6 to retrieve a morpheme (correction word) registered as an error. It is.

Means 4 for judging adjacent character type and acquiring same character type
Is to determine the attribute of the adjacent word of the proofreading word and to combine those of the same character type. The detection result output means 5 performs, for each of the combined words, a different level to highlight and display a replacement word.

Next, the operation will be described. Morphological analysis means 2
Is morphologically analyzed to divide the input sentence into morphemes, morphemes (correction words) of errors registered in the morphological analysis dictionary 6 are extracted, and error levels of the proofwords extracted by the detection result output means 5 are classified. To display highlighting and replacement candidates.

The morphological analysis means 2 morphologically analyzes the input sentence and divides it into morphemes.
The morpheme (correction word) of the error registered in the above is extracted, and the adjacent word having the same attribute in the sentence is synthesized with respect to the proofread word extracted by the adjacent character type determination / identical character type acquisition means 4, and the detection result output means 5 synthesizes it. For the latter word, error levels are classified and the highlighting and replacement candidates are displayed.

Therefore, a morphological analysis dictionary 6 is prepared, an erroneous morpheme (correction word) is extracted from the input sentence, and adjacent words having the same attribute are synthesized. Then, the levels are classified and emphasized according to the error level of the word. By displaying the displayed indications and replacement candidates, it is possible to provide appropriate support for grammar correction, reduce the load of registering words such as katakana, and suppress excessive indications.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, an embodiment and operation of the present invention will be sequentially described in detail with reference to FIGS.

FIG. 1 shows a system configuration diagram of the present invention.
In FIG. 1, an input unit 1 is used to input a sentence, such as inputting a sentence from a keyboard, inputting a proofreading (editing) sentence from an application, and the like.

The morphological analysis means 2 divides the sentence input by the input means 1 into morphemes with reference to the morphological analysis dictionary 6, and extracts the morpheme in which the proofreading information is registered and the proofreading information. Things.

The proofreading information acquiring means 3 includes a morphological analyzing means 2
To obtain the proofreading information of the morpheme extracted from the document. The adjacent character type determination / identical character type acquisition unit 4 determines the character type of the adjacent word in the sentence of the morpheme (correction word) extracted by the morphological analysis unit 2 and extracts and synthesizes the same character type.

The detection result output means 5 is for categorizing the detected words (correction words) to highlight the words and displaying replacement words. The morphological analysis dictionary 6 is a dictionary necessary for performing morphological analysis, and includes information (word notation, part of speech, meaning, grammatical information) necessary for analysis in association with words, and proofreading information ( The correct morpheme, the fact that the morpheme is correct, etc.) are registered in advance.

The character type determination table 7 registers character types and uses them at the time of determination (see FIG. 3). FIG.
4 shows an example of a morphological analysis dictionary of the present invention. As shown in the figure, the dictionary 6 for morphological analysis stores, in advance, proofreading information in association with words (morphemes), word notation (not shown) in the case of Chinese characters, part of speech, meaning, grammatical information, and the like. It is.

The proofreading information indicates the proofreading information of the registered word. When the proofreading information is correct, the proofreading information "None" is registered. When the registered word is wrong, the correct word is registered.

FIG. 3 shows an example of a character type determination table according to the present invention. The character type determination table 7 is a table in which the character types as shown (Katakana, Hiragana, alphabets, Chinese characters, numbers, etc.) are registered in advance as shown below.

Character type: Katakana A ... Ken Hiragana A ... Ken Alphabet A ... Z Z Kanji ... Number 0, 1, 2, 9 Next, the operation of the configuration of FIG. 1 will be described in detail according to the order shown in the flowchart of FIG.

FIG. 4 is a flowchart illustrating the operation of the present invention. In FIG. 4, S1 inputs a sentence. For example, enter the sentence “Computer wheels that have become a hot topic in recent years”.

In step S2, the morphological analysis is performed with reference to the morphological analysis dictionary 6. That is, the sentence input in S1 is subjected to morphological analysis while referring to the morphological analysis dictionary 6, and divided into morphemes as indicated by /. Here, the underline is shown in FIG.
Indicates that the proofreading information (correct morpheme “virus” in the example of this morpheme “Virus”) is registered in the morphological analysis dictionary 6.

"Recent / Topic / Nii / N / N / N / Computer / Beals / Ha " S3 extracts a morpheme. In this method, the morphemes subjected to morphological analysis in S2 are extracted one by one from the beginning.

In step S4, it is determined whether or not error correction information is registered in the morpheme extracted in step S3. For morphemes (correction words) with erroneous proofreading information, the processing after S5 is executed. On the other hand, a morpheme without calibration information is extracted in S3 for the next morpheme without performing the processing in S5 and thereafter, and the determination in S4 is repeated.

By the above-described processing of S1 to S4, morphemes are extracted from the sentence, and only morphemes (correction words) in which error correction information is registered for the extracted morphemes are extracted, and the process proceeds to the next S5 and subsequent steps. Other morphemes are S5
It is excluded here from the following targets.

In step S5, adjacent words (morphemes) having the same attribute are synthesized. That is, for morphemes having error correction information, the adjacent words (adjacent morphemes) in the text are combined when they are of the same character type with reference to the character type determination table 7. For example, in the above example, since the adjacent morpheme “computer” is the same character type in katakana for the morpheme “biels”, the morpheme “computer” is synthesized to be “computer biels”.

In step S6, the result of the synthesis is determined. here,
It is determined whether the synthesis result is “unregistered word + proofreading word” or not. In the case of “unregistered word + proofreading word”, an error level 3 is indicated in S7. For example, “Computer Beams” is a composite word (composite morpheme) of “Compiler” and “Beals”. The former is not registered in the morphological analysis dictionary 6 of FIG. Since the “proofreading word” is registered and pointed out as an error, the synthesis result is “unregistered word + proofreading word”. On the other hand, in other cases of S6, the process proceeds to S8.

In step S8, the attribute is determined. Here, it is determined whether the attribute is “only the proofreading word” or other. For example, since "Beales" is registered as a proofreading word alone in the morphological analysis dictionary 6 in FIG. 2, it is pointed out as an error level 1 (a level with the highest error strength) in S9. On the other hand, in other cases of S8 (other than the proofreading word), it is indicated as the error level 2. For example, "computer wheels" means that "computer" and "beams" are registered with reference to the morphological analysis dictionary 6 in FIG. 2, but are not proofreading words because no error is pointed out, and the latter are registered. Since the error has been pointed out and an error has been pointed out (because the correct "virus" has been registered), the result is other than (other than the proofreading word alone).

As described above, the input sentence is subjected to morphological analysis to obtain a proofreading word (word indicating an error) registered in the morphological analysis dictionary 6, and adjacent words of the proofreading word having the same type are read out. In the case of, combining is performed, the combining result is determined, and the attribute is determined. Error levels 1, 2, and 3 are sequentially determined in the order of strong error, and highlighted, and replaced in accordance with the highlighted display. Candidates (e.g., candidates for replacing the correct word "virus" of the proofreading word "Beals" in FIG. 3) are displayed, and when a replacement candidate is selected, the proofreading word is automatically replaced with the correct proofreading word to proofread the correct word. .

FIG. 5 shows an example of a result indicated by the present invention. The left column in this example of the indication result is “input word (morpheme)”, the center column is “indication range” (calibration information extracted from the morphological analysis dictionary 6 in FIG. 2), and the right column is “ Indication level "(error levels 1, 2, and 3 determined by the flowchart of FIG. 4).

For example, the input word “Viels” in the upper row
Is that the correct word from the morphological analysis dictionary 6 in FIG. 2 is “virus”, and the level 1 in S9 in the flowchart in FIG.
Is determined. The same applies hereinafter. In addition, indication level 1,
Reference numerals 2 and 3 indicate the following.

Pointing level 1: Pointing out the relevant pointing range notation and returning a paraphrase candidate (strongly pointing out as an error word) Pointing level 2: Only pointing out the relevant pointing range notation (pointing out as a word that may have an error) ).

Pointing level 3: Pointing out the corresponding pointing range notation only (pointing out as a word having a low possibility of error, but as a word which may not be a general description) FIG. 6 shows a screen of the present invention. The display example is shown. This screen display example, co-pewter Byi Angeles that is a word in recent years the topic that has been input is to highlight the portion of the underline of the two-stage [Break] "computer Byi Angeles", "check the number at the bottom : 1 error level 2: computer virus "(the number of checks is the number of checked items, error level 2 is the level of error, and subsequent" computer virus "is a replacement candidate. When this replacement candidate is selected, Automatically edits (automatic proofreading), replacing with the automatically entered word "computer wheels"
You can do it.

[0032]

As described above, according to the present invention,
A morphological analysis dictionary 6 is provided to extract an erroneous morpheme (correction word) from an input sentence, synthesize morphemes having the same attribute, and display highlighting and replacement candidates according to the error level of the word. Because we adopt
In addition to providing appropriate support for grammar correction, it is possible to reduce the burden of registering words such as katakana and suppress excessive pointing.

[Brief description of the drawings]

FIG. 1 is a system configuration diagram of the present invention.

FIG. 2 is an example of a morphological analysis dictionary of the present invention.

FIG. 3 is an example of a character type determination table according to the present invention.

FIG. 4 is a flowchart illustrating the operation of the present invention.

FIG. 5 is an example of an indication result of the present invention.

FIG. 6 is a screen display example of the present invention.

[Explanation of symbols]

 1: input means 2: morphological analysis means 3: proofreading information acquisition means 4: adjacent character type determination / identical character type acquisition means 5: detection result output means 6: morphological analysis dictionary 7: character type determination table

Claims (5)

[Claims] 1. A notation indicating device for indicating an error in the notation of a character string in a sentence, comprising: means for morphologically analyzing a sentence to extract an error morpheme registered in an analysis dictionary; A notation indicating device, comprising: means for dividing and emphasizing the level. 2. A notation indicating device for indicating an error in the notation of a character string in a sentence, comprising: means for morphologically analyzing the sentence to extract an erroneous morpheme registered in the analysis dictionary; A notation indicating device comprising: means for synthesizing adjacent words having the same attribute therein; and means for highlighting divided morphemes at different error levels. 3. The method according to claim 1, wherein the level is divided and highlighted, and replacement candidates are also displayed.
Alternatively, the notation pointing device according to claim 2. 4. A computer-readable recording device which records a program for functioning as a means for morphologically analyzing a sentence and extracting error morphemes registered in an analysis dictionary, and a means for dividing and highlighting the morphemes extracted for error levels. Possible recording medium. 5. A means for morphologically analyzing a sentence to extract an erroneous morpheme registered in the analysis dictionary; a means for synthesizing adjacent words having the same attribute in the sentence with respect to the fetched morpheme; A computer-readable recording medium in which a program for functioning as a means for highlighting morphemes at different error levels is recorded.