KR20060111920A - System or method of detecting mis-written word and supplying information to correct it - Google Patents

Abstract

A system and a method for offering miswritten word detection and correction information are provided to generate and present the miswritten word detection/correction information to a user by checking a word including a spelling error and the word which is misused in the context. A miswritten word database(2) stores frequent miswritten words. A word instance database(4) stores correct instances of the words stored in the miswritten word database. A word obtaining module obtains the words from writing. A miswritten word determining module(1) determines the miswritten word by searching the word in the miswritten word database. An instance extracting module(3) extracts the correct instance of the miswritten word by searching the miswritten word from the word instance database.

Description

Determination of error suspects and providing correction information system and its implementation method {System or Method of Detecting Mis-Written Word and Supplying Information to correct it}

1. Error suspect word discrimination module 2. Advanced error word information database

3. Usage Extraction Module 4. Word Usage Database

5. Correction word extraction rule 6. Correction word database

As we enter the information society, prudence is disappearing when people use language. Rather than spelling out the proper vocabulary, the spelling is frequently missed because it is often used only in vocabulary, and it is pronounced as it is pronounced. The feeding field has become a social problem, and various quizzes and culture programs for the proper use of language on TV and radio are increasing rapidly.

Word processors used to compose documents through a computer include a spell checker, but when there is no analysis result in which the morphemes are perfectly combined with the parts-of-speech combination rule by morphing the word, the word is converted into an error word. I use the way of thinking. However, since this method does not take meaning into account, it is not uncommon to find morphemes that cannot be semantically combined because they are misinterpreted, so that the error word is misinterpreted and cannot be found as an error word.

Moreover, the word itself is not an error word, but it does not find any error word that is not contextually correct in the sentence used. Example sentence

"Throwed the stone. And ran away."

In 'and' is not an error word in itself, but the word 'and' is misused where it should be used.

The present invention was devised in consideration of the above points, and is a method of implementing a program module that can be called and used by various application programs such as a word processor that requires the discovery of incorrectly used words. The purpose is to provide an error detection word detection method that detects an improperly used word, and a method of providing correction information to the user by generating appropriate correction information.

Error heartbeat detection method according to the present invention for achieving the above object,

Error information containing a word that is morphological but can be an error, an error label indicating that the word is an error, a word that can be an error depending on the context, and normal context information when the word is used in a normal context. A database is available,

(1) receiving a target word to determine whether there is an error and a context text using the word;

(2) extracting candidate words from the words;

(3) searching the candidate phrase in an error information database;

(4) determining that the word is a suspected error word when there is information indicating that the error word is found as a result of the search;

(3) extracting contextual information from the contextual text, when normal contextual information exists as a result of the search;

(5) comparing the normal context information with the context information;

And (6) if the normal contextual information and the contextual information are different from each other, determining the word as an error suspected word.

It will be described in detail with reference to the drawings.

1 is an error doubt clause determination system according to the present invention and includes an error doubt clause determination module 1 and an error information database 2.

The error suspect word discrimination module receives a word and a context text using the word from a general computer application program requiring a text input, and retrieves the information obtained by searching the error information database 2 and the context information extracted from the context text. It is determined whether the word suspected word error of the word using.

The error information database 2 is constructed of information obtained from a large corpus through natural language processing technology, or is prepared by language experts by hand.

2 is a flowchart of a method for determining an error question clause according to the present invention.

S10 is a step in which the error suspect word discrimination module receives the word and the context text in which the word is used from a computer application program that needs to determine whether a word is used in a context or that there is no spelling error.

The context text may be a sentence or paragraph including the word or a predetermined number of words or sentences located before and after the word.

S20 is a step of extracting candidate words to be searched in the error information database from the words.

In Korean, a word consists of one or more morphemes. Thus, the morpheme in which an error exists can be combined with various other morphemes to generate a number of error words. In this case, it is inefficient to keep all the error phrases in the error information database. Therefore, it is preferable to extract and search for an error candidate word separately from the word. For example, in the case of the word 'in the kitchen' which is incorrectly spelled 'in the kitchen', 'in the kitchen', 'in the kitchen' and 'the kitchen' are extracted as candidate words.

S30 is a step of searching for the candidate word in the error information database.

In the error information database, candidate words as in s20 extracted from frequently occurring error words and error words are extracted and stored as index words. In addition, the error word and the index word generated by the candidate word are stored together with an error mark indicating information indicating that an error. In addition, the words used in contexts that are not errors but are not suitable for themselves, and the candidate words extracted from these words are also stored in Sacker, and the context information extracted from the context text using the words is normal context information. It is stored in the database.

S40 checks whether the error label information exists when the error information database search is successful. If the error label information is present, the error suspect word search target word is determined as an error suspect word (s90).

S50 is a step of checking whether the normal context information exists when the error information database search is successful and the error cover information does not exist. If the normal context information does not exist, it is determined that it is not an error suspect clause.

S60 is a step of extracting context information from the context text to confirm whether the word is normally used in the context text when the normal context information exists.

The context information is information about a context in which the word is used, including a vocabulary used with the word in the context text, a part of speech of the word, and a sentence structure. The vocabulary may be extracted within the word.

S70 is a step of comparing the context information with the normal context information.

And comparing the context information extracted from the context text using the word with the normal context information extracted from the context text when the word is normally used.

If the contextual information and the normal contextual information are different (s80), it is determined that the word is likely used properly in the contextual text (s90) and is determined as an error suspected sentence (s90).

Of course, it is practically difficult for the error information database to store all the normal contextual information necessary for each error word. However, by extracting the normal context from the frequently occurring contexts, the accuracy of error suspect clause determination is increased.

In addition, it is possible to improve accuracy by adding a step of comparing the similarity by extracting abnormal context information from the inappropriately used word and the context text in which the word is used.

Another more improved embodiment of the present invention is shown in Figure 3.

For more effective use of the present invention, it is possible to provide a user with an example of providing corrected word information for the error suspected phrase, an example of an article correctly used with the error suspected phrase, and an example of an article correctly used with the corrected phrase.

To this end, a corrected word database (6) that collects corrected words for words in the error information database (2), and a word example database (4) that contains examples of the phrases and the correct words used in the correct words It is provided.

The corrected word extraction module 5 checks whether the corrected word clause exists in the corrected word clause database 6 for the word determined by the error suspected word discrimination module 1 as the error suspected word clause. 3) The word usage example database (4) is searched for whether there is a usage example of the sentence in which the phrase and the correction phrase are used correctly.

For example, the word `` in the kitchen '' is used to inform the user that the word is in question, and the example includes `` Mother cooks in the kitchen '' with the correction word `` in the kitchen. '' .

As another example, in some of the articles 'Teachers of Students', the word 'pointing' of the error word is a 'teacher who teaches the students' or 'direction' indicating the word to the user. Use examples such as' Of course, one or more examples of writing may be provided.

Herein, the terms or words used in the present specification and claims are not to be construed as being limited to the ordinary or dictionary meanings, and the inventors appropriately define the concept of terms in order to explain their own invention in the best way. It should be interpreted as meaning and concept corresponding to the technical idea of the present invention based on the principle that it can. Therefore, the embodiments described in the specification and the drawings shown in the drawings are only the most preferred embodiment of the present invention and do not represent all of the technical idea of the present invention, various modifications that can be replaced at the time of the present application It should be understood that there may be equivalents and variations.

According to the present invention, an improperly used error that could not be detected using a conventional spelling error checker is easily found to provide an error in a word or spelling error, thereby providing a correct language using environment.

Claims (8)

As a method of determining whether a word is likely to exist, Words that can be stemmed but are errors, error label information indicating that the word is an error, and words that may be errors depending on the context, and error information containing normal context information when the word is used in a normal context. A database is available, (1) receiving a target word to determine whether there is an error and a context text using the word; (2) extracting candidate words from the words; (3) searching the candidate phrase in an error information database; (4) determining that the word is a suspected error word when there is information indicating that the error word is found as a result of the search; (3) extracting contextual information from the contextual text, when normal contextual information exists as a result of the search; (5) comparing the normal context information with the context information; And (6) if the normal context information is different from the context information as a result of the comparison, determining the word as an error suspect word. As a method of determining whether a word is likely to exist, Contains a word that is morphological but can be an error, error label information indicating that the word is an error, and abnormal context information, which is a word that can be an error depending on the context, and contextual information extracted from the context in which the word is inappropriately used. Has a database of error information (1) receiving a target word to determine whether there is an error and a context text using the word; (2) extracting candidate words from the words; (3) searching the candidate phrase in an error information database; (4) determining that the word is a suspected error word when there is information indicating that the error word is found as a result of the search; (3) extracting context information from the context text when abnormal context information exists as a result of the search; (5) comparing the abnormal context information with the context information; And (6) if the abnormal context information and the context information are similar, determining the word as an error suspect word. The method according to claim 1 and 2, The candidate word is the word itself; and And a word that is a remainder except for possible investigation or endings from the word. The method according to claim 1 and 2, And the context information includes a vocabulary extracted from the context text. As a system that finds words that people who write, often use inappropriately, and provides users with examples of the words that are used correctly, Error suspect word database that stores words that are often inappropriately used; A word usage example database storing usage examples of a sentence in which the words stored in the error suspect word database are correctly used; A word acquisition module for obtaining a word from the written text; An error syllable clause determination module that searches for the word in the error syllable clause database and determines whether the word has an error symptom clause; A usage extraction module for retrieving a word determined as an error suspect word from the word usage example database and extracting an example of a sentence in which the word is appropriately used; And And display means for expressing an example of the extracted text to a user. A system that detects an error word and provides a user with a correction word and a usage example of the sentence correctly used, An error suspect word database storing frequently occurring error words; A correction word phrase database for storing correction words of words stored in the error suspect word phrase database and an example of a sentence in which the correction word clause is correctly used; A word acquisition module for obtaining a word from the written text; An error syllable clause determination module that searches for the word in the error syllable clause database and determines whether the word has an error symptom clause; A usage extraction module for retrieving a word determined as an error suspect word from the correction word database and extracting a usage word of the sentence appropriately used for the word and the correction word; And And display means for expressing the extracted correction word and the usage example of the text to a user. The method according to claim 5 or 6, And means for informing a user of a word determined as the suspected error word. The method according to claim 5 or 6, The error suspect clause database is an error information database provided in claim 1 or 2, And the error suspected word determination module uses the error suspected word determination method provided in claim 1 or 2.

Images