JPH06124279A - Text preparation supporting device and method - Google Patents

Abstract

PURPOSE:To provide a correction candidate offering text preparation supporting device and method by detecting the usage error of the homonym, synonym, and near-synonym of a word. CONSTITUTION:This device is equipped with a test word designating part 1 which designates and stores one part of an input text, text analyzing part 2 which detects and stores the related word of the test designated word and a modification relation between the test designated word and the related word, instance dictionary part 4 which stores a word notation, the related word notation, and the modification relation description by making them to correspond to each other, and word dictionary part 5 which stores more than one word dictionary selected from among a homonym dictionary, synonym dictionary, and near-synonym dictionary. And also, the device is equipped with a test designated word usage error discriminating means 3 which discriminates the presence or absence of the usage error of the test designated word from the instance retrieved from the instance dictionary part 4, and the analyzed result stored in the text analyzing part 2, correction candidate instance retrieving means 7 which retrieves the instance from the instance dictionary part 4 about the correction candidate retrieved from the word dictionary part 5, and correct instance correction candidate offering means 8 which detects the instance matched with the analyzed content of the text analyzing part 2, and offers a word notation corresponding to the correct instance.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a text creation support device and a text creation support method, and in particular, automatically detects homonyms, synonyms and synonyms from the input text,
The present invention relates to a text creation support device and method for displaying correction candidates.

[0002]

2. Description of the Related Art As a conventional automatic error detection / correction device for Japanese sentences, there is JP-A-3-88066. The conventional technique will be described below with reference to FIG. When the input Japanese sentence has a word having a homonym in the input Japanese sentence, the conventional sentence creating device detects the word as a test target word and detects a clause that modifies the test target word. Then, whether or not the test target word and the modified phrase of the word have a semantic connection relationship is tested using a semantic connection dictionary that stores the semantic connection availability information between words. Then, if it is determined that there is no semantic concatenation, it is determined that the homonyms are incorrect, and the homonyms of the test target words are searched, and the homonyms that have been searched are compared with the modifier clauses of the test target words. Whether or not there is a concatenation relation is tested using the semantic concatenation dictionary as in the above. As a result, when it is verified that there is a semantic concatenation, the homonyms are presented as correction candidates.

In this way, the above-mentioned conventional technique has realized to automatically detect a usage error of a homonym in a sentence and present a correction candidate.

[0004]

However, the above-mentioned prior art has the following problems. That is, (1) All words having homonyms were used as the test target words, so that even words having no apparent error in the meaning were also tested. (2) Also, because the test object was limited to the usage of homonyms, the usage of synonyms and synonyms should be detected, and the synonyms and synonyms should be included in the search target and used correctly in the text. Could not present possible correction candidates for. (3) Further, in the above-mentioned conventional technique, the operator arbitrarily designates a specific word in the sentence to determine whether or not there is a mistake in using the word, and the operator selects a search dictionary according to the convenience of the sentence creation. Then, I could not use it to specify the search range of correction candidates.

In view of the above situation, the present invention automatically detects not only homonyms but also misuses of synonyms and synonyms for a word designated by an operator in a sentence or an arbitrary word, and a correction candidate is automatically detected. It is an object of the present invention to provide an excellent text creation support device and text creation support method that are excellent in the flexibility and convenience of presentation and that can present accurate correction candidates.

[0006]

In order to solve the above-mentioned problems, the invention of claim 1 specifies a part of an input sentence as a test target word, and stores a notation of the specified test-specified word. The word designation part and the input sentence are analyzed to detect the related words of the test designated word and the modification relation between the test designated word and the related word, and the notation of the detected related word and the description of the modified relation are described as the test designated word. And a sentence analysis unit that stores the word notation and its related word notation and their modified relation description in association with each other,
A word dictionary unit that stores one or more word dictionaries selected from a homonym dictionary, a synonym dictionary, and a synonym dictionary;
An example of the test-specified word is searched from the example dictionary section, and a test-specified word use error determination means for determining whether or not the test-specified word is used incorrectly based on the searched example and the analysis content stored in the sentence analysis section, and the test specification When it is determined that the word has a usage error, a correction candidate is searched from the word dictionary unit, and a correction candidate example searching unit that searches an example from the example dictionary unit for the searched correction candidate, and the searched example is the sentence. An example of corrective example presenting means for detecting an example of matching as a corrective example by comparing with the analysis content stored in the analysis unit and presenting a word notation corresponding to the corrective example as a corrective example correction candidate. And

According to the invention of claim 4, in the invention of claim 1, 1 is selected from the word dictionaries stored in the word dictionary part.
Alternatively, a correction candidate search dictionary designating unit capable of designating two or more word dictionaries as the correction candidate search dictionary is added, and when the correction candidate example searching means determines that the test designation word has a use error, the correction candidate A correction candidate is searched from the word dictionary designated by the search dictionary designation unit, an example is searched from the example dictionary unit for the retrieved correction candidate, and the correct example correction candidate presenting means associates it with the correct example correction candidate. The present invention is characterized in that the word dictionary name for which the correct example correction candidate is searched is presented.

According to the invention of claim 8, in the invention of claim 4, the test word designating part decomposes the input sentence into word units and stores the decomposed word string, and the word decomposing part decomposes. It is characterized by comprising a test word output section for sequentially outputting the notation of each of the words as a test designated word.

[0009]

With the above structure, the present invention has the effect of searching a wide range of correction candidates for supporting text creation and presenting only correct correction candidates. That is, in the invention of claim 1, the test word designating section accepts the designation of the word to be tested and stores it as the test designated word in association with the input sentence.

The sentence analysis unit performs sentence analysis on the basis of the input sentence and the test-specified word to identify a related word to the test-specified word, and also analyzes the modified relation between the test-specified word and the related word to write the related word and It is stored as a modification relation description. The search dictionary unit stores a word dictionary for searching for correction candidates. In addition, the example dictionary unit stores, in advance, for various sentences, a specific word in the sentence, its related word, and their modified relationship in the form of word notation, related word notation, and modified relationship description.

The test-specified word use error determination means retrieves all the examples of the test-specified word from the example dictionary section, and compares all the retrieved examples with the analysis contents analyzed by the sentence analysis section,
It is verified whether or not there is an example that matches the analysis content.
When there is a matching example, it is determined that there is no usage error in the test designation word, and when there is no matching example, it is determined that there is a usage error in the test designation word.

By the above-described operation of each constituent part, the invention of claim 1 determines whether or not there is a usage error only for a specific word arbitrarily designated in the sentence by the operator, and determines whether there is a usage error. In the case of judgment, correction candidates are searched from the word dictionary section, and the correction candidates are narrowed down to appropriate correction candidates by collating the examples with the verification designated word analysis contents and presented as correct example correction candidates. Play.

In addition to the above, the invention of claim 4 searches for a correction candidate from a word dictionary (synonyms, synonyms and synonyms) selected and designated by an operator, and an example of the correction candidate is used. By collating with the contents of the analysis of the test designated word, it is narrowed down to an appropriate correction candidate, which is used as a corrective example correction candidate, and is presented in a form in which the searched word dictionary name is added to the corrective example correction candidate.

According to the invention of claim 8, the word decomposing unit decomposes the input sentence into words and stores the decomposed word string. Further, the test word designating unit stores the notation of each word decomposed by the word decomposing unit and sequentially outputs it as a test designating word. Therefore, the above-mentioned claim 4 for each of the decomposed words.
By repeating the same process as described in,
It has the effect of automatically detecting all misused words in the input sentence and associating the correct example correction candidate with the dictionary section name searched for and proposing it.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The text creation support device of the present invention will be described based on the first and second embodiments. [First Embodiment] FIG. 1 is a block diagram of the text creation support device in the first embodiment.

A test word designating unit 1 designates a part of an input sentence as a word to be tested based on an operator's instruction and stores the designated test designating word. The test word designating unit 1 is composed of a display device, a cursor, a memory and the like. Here, the designation of the test designation word is performed, for example, by the operator designating the first character and the last character of the desired word notation while looking at the text notation on the display. Specifically, if the input sentence is "Today is bright weather", and the operator wants to specify the word "weather clear" as the test designation word, "ten" as the first character and the end By instructing "Akira" as a character and operating the execution key to complete the instruction, the word "weather clear" is designated and stored as the test designation word.

Reference numeral 2 denotes a sentence analysis unit, which analyzes the input sentence based on the test-designated word designated by the test-word designating unit 1, and describes the related words of the test-designated word and the modification between the related word and the test-designated word. Detect and store relational descriptions. Here, the input sentence is analyzed by morphological analysis and syntactic analysis, etc. Even if the test designated word notation and the related word notation are not continuous in the sentence notation, the related word notation of the test designated word and this relation It is possible to detect the modification relation description between the word and the test-specified word.

The content of the sentence analysis will be shown by a specific example. When the input sentence is “indicating a model” and the test designation word is “in a model”, the sentence analysis unit 2 “shows” as the related word notation and “<search designated word> as << related word >> as the modification relation description. Is detected. In addition, the input sentence is a test designation word (model) such as "clearly show model" or "show model by itself"
If there is a notation (clearly or naturally) that is not directly related to the modified relationship between the two and the related word (shown), and the test designation word is “exemplary”, the detection result is It will be the same as the former.

The above description of the modification relation defines the dependency relation between the test designation word notation and the related word notation, and is used in the present invention to detect misuse of the test designation word. In addition, in the above-mentioned specific example “<test designation word> is << related word>”, the dependency relation is established by using the particle “wo” between the test designation word “exemplary” and the related word “show”. Is shown.

Reference numeral 3 is a use error determining means for determining whether or not the test designated word has a use error. The use error determination means 3 is composed of an example retrieval unit 3a and a use error determination unit 3b. The example retrieval unit 3a retrieves all examples of test designation words from the example dictionary unit 4 based on the designation of the test word designation unit 1.
Then, the usage error determination unit 3b collates the example searched by the example search unit with the related word notation and the modification relation description (hereinafter referred to as the sentence analysis content) detected by the sentence analysis unit, and whether there is a match. To test. And if there is a matching example, it is determined that there is no usage error in the test designated word,
On the other hand, when there is no matching example, it is determined that the test designation word has a usage error, and the information is transmitted to the correction candidate search unit.

Reference numeral 4 denotes an example dictionary section, which associates word notations with related words and their modification relationships for a large number of unspecified words and stores them as examples. Here, the example is obtained by performing morphological analysis and structural analysis on a linguistically correct sentence in advance to obtain the word notation in the sentence and its related word notation and their modified relationship, and the word notation, The related word notation and the modifier relations thereof are configured as constituent elements. In the example dictionary unit 4, one block is formed corresponding to the word notation as shown in FIG. The description is stored.

A word dictionary section 5 stores one or more word dictionaries selected from a homonym synonym dictionary, a synonym dictionary and a synonym dictionary. The word dictionary unit 5 is used by the correction candidate search unit 7a to search for correction candidates, and each dictionary is stored in association with a word notation. FIG. 3 illustrates the contents of the homonym synonym dictionary, FIG. 4 illustrates the synonym dictionary, and FIG. 5 illustrates the synonym dictionary.

A correction candidate search dictionary designating unit 6 designates one or more word dictionaries among the word dictionaries stored in the word dictionary unit by the operator as a correction candidate search dictionary. The designation information of the correction candidate search dictionary designation unit 6 is passed to the correction candidate search unit 7a. Reference numeral 7 is a correction candidate example searching means, which searches for a correction candidate and its example. The correction candidate search unit 7 is composed of a correction candidate search unit 7a and a correction candidate example search unit 7b, and the correction candidate search unit 7a receives from the use error determination unit 3 the determination information that the test designation word has a use error. In this case, the correction candidate is searched from the word dictionary unit 5 designated by the correction candidate retrieval dictionary designating unit 6. Then, the correction candidate example retrieval unit 7b retrieves an example from the example dictionary unit 4 for the retrieved correction candidate.

Reference numeral 8 is a corrective example correction candidate presenting means, which detects an appropriate corrective candidate based on the corrective example and presents the result. The corrective example correction candidate presenting means 8 is composed of a corrective example detecting unit 8a, a corrective example correction candidate presenting unit 8b, and a word dictionary name presenting unit 8c, and the corrective example detecting unit 8a includes a corrective candidate example searching unit 7b. The searched example is compared with the sentence analysis content to detect a matching example as a correct example, and the word notation of the correction candidate corresponding to this correct example is passed to the correct example correction candidate presenting section 8b as a correct example correction candidate. The word dictionary name for which the correct example correction candidate is searched is passed to the word dictionary name presenting unit 8c. The correct use example correction candidate presenting unit 8b and the word dictionary name presenting unit 8c present the correction candidate and the word dictionary name in a form corresponding to the correct use example correction candidate as correction information.

The basic configuration of the text creation support device of the first embodiment of the present invention is as described above. Next, the basic operation based on the above configuration will be described with reference to the follow chart of FIG. In addition, a supplementary explanation will be given by giving a specific example in which "I apologize for the usual situation" is specified as the input sentence and "usually" is specified as the test designation word. First, when a sentence is input to the sentence creation support device (s1) and the operator operates the test word designating unit 1 to designate a word to be a test target, the test word designating unit 1 selects the designated word. The notation is stored as a test designation word (s2).

On the other hand, when the test designated word is not designated, the test word designation unit 1 waits for designation of the test designated word. The test designation word is specified by the operator pointing the beginning and end characters of the word to be designated with the cursor while looking at the input text on the display, and pressing the test word designation execution key on the key board. . Specifically, "normal" is specified by instructing "normal" as the first character and "dan" as the terminal character.

When the test designated word is designated, the sentence analysis unit 2 analyzes the input sentence based on the test designated word, and tests the modification relation between the words before and after the test designated word and the modification relation, The description of the modification relation between the word notation and the test designation word and the related word is detected. Then, the detection result is stored (s3). In the above-described specific example, the word "goshinsha" is detected as the related word and the "related word" of "the test specified word" is detected as the related word with respect to the test specified word "usually", and these notations are detected by the sentence analysis unit Will be remembered.

In the following description, the related word notation of the test designated word and the description content of "<< related word of << test designated word >>" will be referred to as "chapter analysis content of test designated word". Then, the example retrieval unit 3a retrieves all examples of the test-specified word from the example dictionary unit 4 and passes them to the usage error determination unit 3b (S4).
In the above specific example, the example search results in, for example, the result shown in FIG.

The usage error determination unit 3b collates whether or not there is a match with the analysis content of the test designated word in the example searched by the example search unit 3a (S5), and if there is a match (S5; Yes). ) Presents to the operator that there is no mistake in using the test designated word (S11). On the other hand, if none of the examples searched in step S5 match the analysis content of the test designated word (s5; No), the usage error determination unit 3b determines that the test designated word has a usage error. The correction candidate searching unit 7
Tell a (s5).

In response to this, the operator operates the correction candidate search dictionary designating unit 6 to search the correction candidates of the test-designated words as word dictionaries, namely the homophone dictionary, the synonym dictionary, and the synonym dictionary. Input 1 or 2 or more to execute designation (S6). The correction candidate search dictionary designation unit 6 waits until the dictionary is designated (S6; N).
O).

By designating the correction candidate retrieval dictionary, the correction candidate retrieval unit 7a retrieves all the correction candidates from the designated word dictionary, and the correction candidate example retrieval unit 7b together with the retrieved dictionary name.
(S7). FIG. 8 shows the search contents of the correction candidate search unit 7a. The correction candidate example retrieving unit 7b retrieves an example from the example dictionary unit 4 for each received correction candidate, and passes all retrieved correction candidate examples to the correct example detecting unit 8a (S
8).

The correct example detecting unit 8a compares each of the correction candidate examples retrieved by the correction candidate example retrieving unit 7b with the analysis content of the test designation word, and when an example matching the analysis content is detected (s9). ; Yes), determine that they are correct examples, and send them to the correct example correction candidate presenting unit 8b. FIG. 9 shows an example of correction candidates that match the analysis contents. The correct example correction candidate presenting unit 8b stores all correct example correction candidates, and the word dictionary name presenting unit 8c receives the searched word dictionary name of the correct example correction candidates from the correct example detecting unit 8a and corresponds to the correct example correction candidates. Let me remember. These data are read and presented as correction information in the form as shown in FIG. 10 according to an instruction from the operator (s10).

If the correct example detecting unit 8a does not find any example that matches the analysis content of the test designated word (s9; No), it is not possible to provide information to the correct example correction candidate presenting unit 8b. Will be displayed (s12). With the above, the sentence creation support in the first embodiment ends. [Second Embodiment] Next, a text creation support device according to a second embodiment of the present invention will be described.

FIG. 11 is a block diagram of the sentence creation support device in the second embodiment of the present invention. In FIG. 11, components having the same functions as those in the first embodiment are designated by the same reference numerals as those in the first embodiment, and the description thereof will be omitted.
A word decomposing unit 101 decomposes an input sentence into words. Reference numeral 102 is a test designated word output unit, which is a word decomposition unit 101.
Each decomposed word is stored and sequentially output to the sentence analysis unit 2 and the example retrieval unit 3a as a test designated word.

The text creation support device of the second embodiment of the present invention is configured as described above. The basic operation is shown in FIG.
It will be described based on the flowchart of FIG. It should be noted that step No. is the same for the same part as the operation of the first embodiment, and the description is omitted. The word decomposing unit 101 decomposes the input sentence into words and sends it to the test designated word output unit 102 (s201). If the sentence of FIG. 12 is input, the word decomposing unit 101 will divide the sentence as shown in FIG.

The test designated word output unit 102 stores each word decomposed by the word decomposing unit 101, sequentially designates it as a test word, and outputs it to the sentence analysis unit 2. In the operation after the above, the operations of steps s3 to s10, step s203, and step s204 are sequentially repeated in the output order of the test word. Here, step s203 determines whether or not the processing has been completed for all the words specified in step s202, and if the processing has not been completed for all the words, the process returns to step s3 to process the next specified target word. I do.

As an example of displaying the correction information in the input sentence shown in FIG. 12, the correction information as shown in FIG. 14 is displayed for each of "negligence", "confidence", and "fellowship". With the configuration described above, the sentence creation support apparatus according to the second embodiment can detect misuse of homonyms, synonyms, and synonyms for an arbitrary word in a sentence, and display the correction candidate.

In the first and second embodiments described above, the correction candidate search dictionary is specified after it is determined in step s5 that there is an error in using the test specification word. However, the correction candidate search dictionary is specified. May be before the correction candidate search unit starts the search for correction candidates in step s7. ,

[0039]

As is apparent from the above description, the sentence creation support apparatus and method of the present invention are: (1) Providing an example dictionary section to determine whether or not there is a mistake in the test designated word based on the searched example. In order to determine whether or not there is a correct usage, it is possible to determine that there is a usage error only for a word that is semantically misused.

Therefore, compared to the prior art, there is an effect that the usage error can be accurately determined. (2) Further, by using the synonym dictionary and the synonym dictionary as the correction candidate search dictionary, it is possible to detect the usage error of the synonym and the synonym and present the correct example correction candidate. Furthermore, since the test word designation unit is provided, it is possible to search the correct example correction candidates within the range of the word dictionary designated by the operator.

Therefore, there is a high degree of freedom in the sentence creation support, and there is an effect that efficient sentence creation support limited to the words required by the operator can be performed. (3) Further, the word decomposition unit decomposes the input sentence into words, and the test designated word output unit sequentially outputs each decomposed word as a test designated word, thereby targeting all the words in the sentence. As a result, the usage error can be automatically determined, and the correction candidate and the dictionary unit name searched for can be associated and presented.

Particularly, in the presenting method of the present invention in which correction information is added to a correction candidate by adding the name of the searched dictionary to the correction candidate, the presence of an incorrect word for the operator and its correct example correction candidate, as well as the incorrect usage, are used. It is also possible to inform the relation between the word that has been corrected and the correction candidate, and it is possible to provide more convenience by creating an error-free sentence and improve the convenience of the sentence creation support.

As described above, the text creation support apparatus and method according to the present invention accurately determines a word usage error for a specified word or all words in a text, and determines an appropriate correction candidate (corrective example correction candidate). According to the present invention, the degree of freedom of the operator with respect to the sentence creation support can be increased, the error word can be detected widely and accurately, and an appropriate correction candidate can be presented. It is possible to provide a text creation support device and a text creation support method that are excellent in practicality.

[Brief description of drawings]

FIG. 1 is a configuration diagram of a sentence creation support device according to a first embodiment of the present invention.

FIG. 2 shows contents of an example dictionary section.

FIG. 3 shows contents of a homonym dictionary unit.

FIG. 4 is the contents of the synonym dictionary section

FIG. 5 shows contents of a synonym dictionary section.

FIG. 6 is a flowchart showing the operation of the text creation support device in the first exemplary embodiment of the present invention.

FIG. 7 shows the contents of an example searched by an example search unit for a test designated word in the first embodiment of the present invention.

FIG. 8 shows contents retrieved by a correction candidate retrieval unit in the first embodiment of the present invention.

FIG. 9 shows the contents detected by the correction candidate detection unit in the first embodiment of the present invention.

FIG. 10 shows the contents of correction information in the first embodiment of the present invention.

FIG. 11 is a configuration diagram of a text creation support device according to a second embodiment of the present invention.

FIG. 12 shows the contents of an input sentence in the second embodiment of the present invention.

FIG. 13 shows a word string decomposed by a word decomposition unit in the second embodiment of the present invention.

FIG. 14 shows the contents of correction information in the second embodiment of the present invention.

FIG. 15 is a flowchart showing an operation of the text creation support device in the second exemplary embodiment of the present invention.

[Explanation of sign]

 1 Test Word Designation Part 1 2 Text Analysis Part 2 3 Usage Error Judgment Means 3 3a Example Search Part 3b Usage Error Judgment Part 4 Example Dictionary Part 5 Word Dictionary Part 6 Correction Candidate Search Dictionary Designation Part 7 Correction Candidate Search Means 7a Correction Candidate Search Part 7b Correction candidate example search unit 8 Corrective example correction candidate presenting unit 8a Corrective example detecting unit 8b Corrective example correction candidate presenting unit 8c Word dictionary name presenting unit 101 Word decomposing unit 102 Test designated word output unit

Claims (16)

[Claims] 1. A test word designating section for designating a part of an input sentence as a test target word and storing the notation of the designated test designated word; and a related word of the test designated word and a test by analyzing the input sentence. A sentence analysis unit that detects a modification relation between a specified word and its related word, and stores the description of the detected related word and the description of the modification relationship in association with the notation of the test specified word, and the word notation and its related word An example dictionary section that stores notations and their modified relationship descriptions in association with each other, a word dictionary section that stores one or more word dictionaries selected from a homonym dictionary, a synonym dictionary, and a synonym dictionary, and a test specification An example is searched from the example dictionary section for words, and a test designated word use error determination means for determining whether or not there is a use error of the test designated word based on the searched example and the analysis content stored by the sentence analysis section, When it is determined that there is a usage error in the constant designation word, a correction candidate example searching unit that searches a correction candidate from the word dictionary unit and searches an example from the example dictionary unit for the searched correction candidate, and the searched example A corrective example correction candidate presenting unit that detects an example that matches the analysis content stored in the sentence analysis unit as a corrective example, and presents a word notation corresponding to the corrective example as a corrective example correction candidate. A text creation support device. 2. The sentence according to claim 1, wherein the corrective example correction candidate presenting means can present, together with the corrective example correction candidate, a word dictionary name in which the corrective example correction candidate is searched for. Creation support device. 3. A correction candidate search dictionary designation unit capable of designating one or more word dictionaries among the word dictionaries stored in the word dictionary unit as a correction candidate search dictionary. The sentence creation support device of item 1. 4. A correction candidate search dictionary designation unit capable of designating one or more word dictionaries among the word dictionaries stored in the word dictionary unit as a correction candidate search dictionary. Item 2 sentence creation support device. 5. A word decomposing unit in which the test word designating unit decomposes an input sentence into words and stores the decomposed word string,
The sentence creation support device according to claim 1, further comprising a test word output unit that sequentially outputs the notation of each word decomposed by the word decomposition unit as a test designated word. 6. A word decomposing unit in which a test word designating unit decomposes an input sentence into word units and stores the decomposed word string,
The test word output unit that sequentially outputs the notation of each word decomposed by the word decomposition unit as a test designated word, and further corrects one or more word dictionaries from the word dictionaries stored in the word dictionary unit. The sentence creation support apparatus according to claim 1, further comprising a correction candidate search dictionary designating unit that can be designated as a candidate search dictionary. 7. A word decomposing unit in which the test word designating unit decomposes an input sentence into word units and stores the decomposed word string,
The sentence creation support apparatus according to claim 2, further comprising a test word output unit that sequentially outputs the notation of each word decomposed by the word decomposition unit as a test designated word. 8. The test word designating unit decomposes the input sentence into word units and stores the decomposed word string, and the notation of each word decomposed by the word decomposing unit is sequentially output as the test designating word. And a correction candidate search dictionary designating unit capable of designating one or more word dictionaries among the word dictionaries stored in the word dictionary unit as a correction candidate searching dictionary. 3. The sentence creation support device according to claim 2. 9. A test word designating step of designating a part of an input sentence as a test target word and storing a notation of the designated test designated word; and a related word of the test designated word and a test by analyzing the input sentence. A sentence analysis step of detecting a modified relation between the designated word and its related word, and storing the notation of the detected related word and the description of the modified relation in association with the notation of the test designated word, the word notation and its related word From the example dictionary section in which the notations and their modified relationship descriptions are stored in association with each other, the test-specified word example search step for searching the test-specified word example and the test-specified word example searched in the test-specified word example search step Based on the analysis contents stored in the sentence analysis step, the test-specified word use error determination step that determines whether or not the test-specified word is used incorrectly, and If it is determined that there, 1 selected homonym dictionary, the synonym dictionary and thesaurus
Alternatively, a correction candidate is searched from a word dictionary unit that stores two or more word dictionaries, and a correction candidate example searching step of searching an example from the example dictionary unit for the searched correction candidate and a correction candidate example searching step are performed. An example in which a matching example is detected as a correct example by collating with the analysis content stored in the sentence analysis step and a word notation corresponding to the correct example is presented as a correct example correction candidate, and a correct example correction candidate presenting step, A text support method characterized by comprising. 10. The sentence support method according to claim 9, wherein the corrective example correction candidate presenting step presents the corrective example correction candidate and the word dictionary name in which the corrective example correction candidate is searched in a corresponding form. . 11. A correction candidate search dictionary designating step for designating one or more word dictionaries among the word dictionaries stored in the word dictionary unit as a correction candidate search dictionary. The sentence support method of paragraph 9. 12. A correction candidate search dictionary designating step for designating one or more word dictionaries among the word dictionaries stored in the word dictionary unit as a correction candidate search dictionary. The sentence support method of paragraph 10. 13. The test word designating step sequentially outputs, as a test designating word, a word decomposing step of decomposing an input sentence word by word and storing the decomposed word string, and a notation of each word decomposed in the word decomposing step. 10. The sentence creation support method according to claim 9, further comprising a test word output step. 14. The sentence support method according to claim 9, wherein the corrective example correction candidate presenting step presents the corrective example correction candidate and the word dictionary name for which the corrective example correction candidate is searched in a form corresponding to the corrective example correction candidate. . 15. A correction candidate search dictionary designating step for designating one or more word dictionaries among the word dictionaries stored in the word dictionary unit as a correction candidate search dictionary. 13 sentence support methods. 16. A correction candidate search dictionary designating step for designating one or more word dictionaries among the word dictionaries stored in the word dictionary unit as a correction candidate search dictionary. 14 text support methods.