KR100311224B1 - Apparatus for determining linking information of english verbal phrase and method using the same - Google Patents

Abstract

The present invention relates to an apparatus and method for determining connection information of English phrases, comprising: a first step of extracting combinations of all phrases using syntax structure information of a Korean sentence; A second step of extracting a connection pattern list for the connection relationship combination of the extracted terms; A third step of determining whether the extracted connection pattern satisfies a Korean connection pattern condition; A fourth step of selecting a basic connection pattern from the connection pattern database according to the determination result of the third step; A fifth step of selecting a corresponding connection pattern according to the determination result of the third step; A sixth step of determining whether all connection patterns to be applied are selected for all of the extracted connection relationship combinations; A seventh step of determining, based on the determination result of the sixth step, the connection pattern set selector to determine the most suitable connection pattern set capable of providing all of the English connection information from the list of selected connection patterns; An eighth step of moving to the second step according to the determination result of the sixth step; And a ninth step of providing connection information and generation information of an English terminology using the connection information of the connection pattern set in the English generation information reflector.

Description

Apparatus and method for determining connection information of English idiom {APPARATUS FOR DETERMINING LINKING INFORMATION OF ENGLISH VERBAL PHRASE AND METHOD USING THE SAME}

The present invention relates to an apparatus for translating a Korean sentence into an English sentence and a method thereof, and in particular, to select an English conjunction corresponding to a connecting ending connecting a Korean terminology using a connection pattern, and to be translated in correspondence with the Korean terminology. An apparatus for determining connection information of an English vocabulary for providing order information and English generating information about English phrases, and a method thereof and a computer-readable recording medium recording a program for realizing the method.

In general, one sentence is generated by syntax rules, and all native speakers have the ability to extend sentences using conjunctions.

Therefore, a large number of Korean sentences to be translated in the Han-Young auto-prosperity system have a connection structure of verbal phrases.

Since Korean has the characteristics of refractions, each phrase is connected using the connecting ending of the word. Therefore, in the English-Korean automatic translation system that performs translations in the unit of phrases, in order to semantically correctly translate the translation results of the English phrases corresponding to each Korean phrase, a proper English conjunction corresponding to the Korean ending is connected. In addition, by determining the relative order between the phrases, by providing the information necessary to generate the English phrases, complete English sentences are generated.

In addition, Korean has not only a variety of connection endings, but a connection ending plays a role in connecting sentences with one or more meanings. The selection of the appropriate English conjunction is not possible.

In the conventional method of determining connection information of English idioms, the English connection information for Korean connection endings is determined only for some idioms, without special treatment for various uses of one Korean connection ending, and in other cases For each linking ending, the most common English conjunctions are used as translation information to carry out the connection of English phrases.

However, the conventional method of determining the connection of English idioms as described above, when the Korean connection ending is not used as the typical usage, provides an inappropriate English conjunction selection and English idiom generation information. There is a problem that the translation fails.

Accordingly, the present invention has been made to solve the above problems, and relates to the connection of English phrases and generation using the Korean language frame for determining the connection pattern and the correct connection pattern for the connection endings connecting the Korean language phrases. An object of the present invention is to provide an apparatus for determining connection information of an English terminology capable of providing translation information, and a method and a computer readable recording medium storing a program for realizing the method.

1 is a configuration diagram of an embodiment of a connection information determination device of the English vocabulary according to the present invention.

Figure 2 is a detailed configuration diagram of an embodiment of a partial band door connector according to the present invention.

3 is a flowchart illustrating an embodiment of a method for determining connection information of an English vocabulary according to the present invention.

* Explanation of symbols for the main parts of the drawings

100: connection information determination device 101: stemmer analyzer

102: division analyzer 103: partial band door connector

104: partial band gate generator 105: band gate refiner

106: stemmer generator 200: dependency extractor

201: Connection Pattern Explorer 202: Connection Pattern Set Selector

203: English generated information reflector 204: connection pattern database

205: Korean Dictionary of War

An apparatus of the present invention for achieving the above object, In the apparatus for determining the connection information of the English phrases applied to the Korean-English translation system, morpheme analysis means for performing a morphological analysis of the Korean sentence to be translated; Syntactic analysis means for performing structural analysis of a phrase unit using the analyzed information of the morpheme analysis means; Dependency relationship extraction means for receiving syntax structure information of a Korean sentence from the syntax analysis means and extracting a combination of all of the phrases connectable through a connection ending; Connection pattern storage means for storing a connection pattern of Korean idioms and basic connection patterns having the most frequent English translation information for each connection ending; Korean frame information storage means for determining a connection pattern to be applied to the Korean sentence to be translated in the connection pattern storage means; Connection pattern searching means for extracting a connection pattern list applicable to the dependency relationship combinations of the phrases extracted by the dependency relationship extracting means using the connection pattern storage means and the Korean language information storage means; Connection pattern set selection means for determining a most suitable connection pattern set that can provide all of the English connection information of connection terms of Korean sentences to be translated among the connection patterns searched by the connection pattern searching means; English generation information reflecting means for providing connection information and generation information of the English language phrases corresponding to the Korean language phrases using the connection information of the connection pattern set selected by the connection pattern set selection means; Partial band sentence generation means for performing an English translation of a short sentence unit for each phrase received from the English generation information rebel means; Band sentence refining means for performing skipped component restoration work and style correction for the English sentence generation, such as sex, number, tense match information and subject for each sentence translated from the partial band sentence generation means into short sentences; and It includes a morpheme generating means for receiving the refined phrases from the band sentence refining means to generate an English morpheme.

On the other hand, the method of the present invention, the method of determining the connection information of the English language phrases applied to the apparatus for determining the connection information of the English language phrases, comprising: a first step of extracting combinations of all the phrases using the syntax structure information of the Korean sentence; Extracting a connection pattern list from a connection pattern database with respect to the connection relationship combination of the extracted phrases; A third step of determining whether a connection pattern extracted from the connection pattern list satisfies a Korean connection pattern condition; A fourth step of selecting a basic connection pattern corresponding to the extracted terminology from the connection pattern database if the Korean connection pattern condition is not satisfied as a result of the determination in the third step; A fifth step of selecting a corresponding connection pattern if the Korean connection pattern condition is satisfied as a result of the determination in the third step; A sixth step of determining whether all of the connection patterns to be applied to the connection relationship combinations of all extracted terms are selected; As a result of the determination in the sixth step, when all of the connection patterns to be applied to the connection relationship combination of all the phrases are selected, the connection pattern set selector for the connection relationship of the phrases of the Korean sentence to be translated from the list of selected connection patterns A seventh step of determining the most suitable set of connection patterns capable of providing all of the English access information; An eighth step of skipping to the second step if the connection pattern to be applied to the connection relationship combination of all the phrases is not selected as the result of the sixth step; And a ninth step of providing, in the English generation information reflector, connection information and generation information of the English language phrase corresponding to the Korean language phrase using the connection information of the selected connection pattern set.

The present invention also provides an apparatus for determining English phrase connection information having a microprocessor, comprising: a first function of extracting combinations of all phrases using syntax structure information of a Korean sentence; A second function of extracting a connection pattern list from a connection pattern database with respect to the connection relationship combination of the extracted terms; A third function of determining whether a connection pattern extracted from the connection pattern list satisfies a Korean connection pattern condition; A fourth function of selecting a basic connection pattern corresponding to the extracted terminology from the connection pattern database if it is determined that the third function does not satisfy the Korean connection pattern condition; A fifth function of selecting a corresponding connection pattern if the Korean connection pattern condition is satisfied as a result of the determination of the third function; A sixth function of determining whether all of the connection patterns to be applied are selected for the connection relationship combinations of all extracted terms; As a result of the determination of the sixth function, if all of the connection patterns to be applied to the connection relationship combination of all the phrases are selected, the connection pattern set selector for the connection relationship of the phrases of the Korean sentence to be translated from the list of selected connection patterns A seventh function of determining a set of most suitable connection patterns capable of providing all of the English connection information; An eighth function of skipping to the second function if the connection pattern to be applied to the connection relationship combination of all the phrases is not selected as a result of the determination of the sixth function; And a program for realizing a ninth function of providing connection information and generation information of an English language phrase corresponding to a Korean language phrase using the connection information of the selected connection pattern set in the English generation information reflector. Provide a record carrier.

The above objects, features and advantages will become more apparent from the following detailed description taken in conjunction with the accompanying drawings. Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a configuration diagram of an embodiment of the apparatus for determining connection information of an English vocabulary according to the present invention.

As shown in the drawing, the apparatus for determining connection information of an English terminology according to the present invention uses a morpheme analyzer 101 for performing a morphological analysis of a Korean sentence to be translated and analyzed information of the morpheme analyzer 100. By using the parser 102 to analyze the structure of the phrase phrase unit, and using the analyzed Korean syntax information to determine the connection relationship between each phrase, the selection of English conjunctions and the word order between the phrases to be translated and generated A partial band sentence connector 103 for providing information, a partial band sentence generator 104 for translating a short sentence unit of English for each phrase, a gender, number, tense match information and subject for generating an English sentence, etc. A band sentence refiner 105 for performing an abbreviated component restoration operation and a style correction operation, and a morpheme generator 106 for generating an English morpheme.

2 is a detailed block diagram of an embodiment of a partial band connector (103) according to the present invention.

As shown in the drawings, the partial band sentence connector 103 to which the present invention is applied is a combination of dependency relations of idioms that can be connected using a connection pattern from syntax structure information of a Korean sentence, that is, connectable through a connection ending. Applicable to the dependency extractor 200 extracted from the dependency extractor 200 using the dependency extractor 200 that extracts a combination of all phrases, and the connection pattern database 204 and the Korean dictionary dictionary 205. The connection pattern explorer 201 extracting the connection pattern list and the English connection information for the connection relations of the phrases of the Korean sentence to be translated among the connection patterns found in the connection pattern explorer 201 can be provided. A connection pattern set selector 202 for determining a set of the most suitable connection patterns, and a connection pattern selected by the connection pattern set selector 202. English generation information reflector 203 which provides access information and generation information of English idioms corresponding to Korean idioms using a set of connection information, and a connection pattern of Korean idioms and the most frequent English for each connection ending. A connection pattern database 204 storing a basic connection pattern having translation information, and a Korean dictionary dictionary 205 for determining a connection pattern to be applied to a Korean sentence to be translated in the connection pattern database 204. do.

In order to determine which role is used among the various usages of the linking ending, the meaning of a plurality of linking endings should be understood.

For example, when translating the sentence "I was very surprised and sad to hear the news," the sentence consists of three idioms, which are connected by the linking word 'Go'. However, the word 'go' in 'listening' and 'go' in 'to be surprised' are connected in the same form as the term 'go', but its usage is different. In other words, 'high' in 'listening' is used as a linking ending indicating a reason, so 'to infinitive' or 'because' is used in the English sentence, but 'high' in 'playing' is a connecting ending meaning equal connection. In conjunction with a conjunction such as 'and' in an English sentence.

As such, the reason why the use of the link word 'Go' of the same word is different is determined by how the meanings of the sentences connected by the link word are connected.

The linking pattern is for the two or more idioms within the range that can determine the conjunctions connecting the English bands of each idiom for two or more Korean idioms connected by a connecting ending, and the information related to each verb and the English conjunction information It is described in terms of, and provides information on English conjunctions used in connection of words, word order information between words, and English generation information of each word phrase.

In addition, the connection pattern is divided into a Korean connection pattern portion describing the conditions under which the connection pattern can be applied, and an English connection pattern portion providing English translation information.

The Korean connection pattern portion describes dependency relations between idioms using parentheses, and includes indexes of idioms for each idiom, identifiers of the frames corresponding to the idioms, and connection ending information of idioms.

And, the English connection pattern portion describes the English idiom corresponding to the index of the idiom described in each Korean idiom. In this case, the dependency relationship between the phrases is not described in parentheses information in the English connection pattern portion, but appears in the English connection pattern in the order in which the English phrases appear in the English sentence.

The connection pattern is applied in the same manner to describe the fixed expression in addition to the connection information on the connection mother, so that the processing of the fixed expression and the processing of the connection mother can be performed simultaneously.

When connected in the form of 'phrase 1 + connective terminal 1 terminology 2', the form of the connection pattern used to select the correct conjunction corresponding to the connecting terminal 1 is as follows. Can be extended to verbs.

(Index of Phrase 1: Index Index, Linked End Information) Index of Phrase 2: Index Index)

=> [Index: generated information of phrase 1] [junk information] [index: generated information of phrase 2]

Now, with reference to the accompanying Figure 3 will be described in detail with respect to the method of determining the connection information of the English vocabulary according to the present invention.

3 is a flowchart illustrating an embodiment of a method for determining connection information of an English terminology according to the present invention.

First, the morphological analysis and syntax analysis are performed on the input Korean sentence (300), and the dependency relations of the phrases that can be connected using the connection pattern from the syntax structure information of the Korean sentence to be translated through the dependency extractor 200. Combination of, that is, the combination of all the phrases that can be connected through the connection ending is extracted (302).

Next, using the Korean dictionary dictionary 205, it is checked whether or not a suitable connection pattern exists in the connection pattern database 204 for the dependency combination of the phrases extracted through the connection pattern explorer 201, and the applicable connection pattern list After extraction (304), it is determined whether the corresponding connection pattern satisfies the Korean connection pattern condition for each connection pattern in the connection pattern list (306).

If the determination result 306 does not satisfy the Korean connection pattern condition, the basic connection pattern is selected from the connection pattern database 204 (308). If the Korean connection pattern condition is satisfied, the corresponding connection pattern is selected (310). ).

In operation 312, it is determined whether all of the connection patterns to be applied to the connection relationship combinations of the extracted terms are selected.

As a result of the determination 312, when all the connection patterns to be applied to the connection relationship combination of all the phrases are selected, the connection pattern set selector 202 connects the relationship of the phrases of the Korean sentence to be translated from the list of the selected connection patterns. A set of most suitable connection patterns capable of providing all of the English access information for is determined (314).

As a result of the determination 312, if all the connection patterns to be applied to the connection relationship combination of all the phrases are not selected, the process of extracting the applicable connection pattern list 304 is performed.

Finally, the connection information and generation information of the English language phrase corresponding to the Korean language phrase is provided using the connection information of the connection pattern set selected by the English generation information reflector 203 (316). In this case, when the Korean sentences to be translated are connected by two or more connection patterns, final information about the final production of English phrases is determined by using dependency information of Korean phrases and relative order information of phrases in a sentence.

In the example sentence `` I was very surprised and sad to hear the news, '' the connection pattern used to determine the connection of English idioms,

(([VP1: Frame1] [High]) [VP2: Frame2]) => [VP1] [and: Equal] [VP2]

(([VP1: Frame2] [High]) [VP2: Frame3]) => [VP2] [to: Reason] Same as [VP2: INF], and the frame information for determining the connection pattern to use is

Battle 1 ['Listen': 'People'-Give' News'-Objects]

Battle 2 ['Amazing': 'People'-Giving]

Same as Battle 3 ['Sad': 'People')

Regarding the example sentence 'I was very surprised and sad to hear the news', the parser 102 performed a syntax analysis of the phrase unit through the syntax analysis (' (I was very surprised and sad)). When the dependency structure of the form is provided, the partial band sentence connector 103 extracts a combination of connection relations among all possible phrases from the syntax structure information, and then the candidate of the extracted phrase connection relation and the connection pattern. Among the connection patterns in the database 204, it is checked whether there is a connection pattern that matches the conditional terms of the Korean terminology of the connection pattern, and selects the optimal connection patterns capable of connecting the entire Korean sentence among the matching connection patterns.

Finally, by selecting the correct conjunctions and providing the English generation information for each connection pattern, the English sentence to be translated becomes '[I'm sad] and [surprise] to [heard the news]'. To decide.

The present invention described above is capable of various substitutions, modifications, and changes without departing from the spirit of the present invention for those skilled in the art to which the present invention pertains. It is not limited to the drawing.

The present invention as described above, by providing the translation information about the connection and generation of English idioms using a Korean pattern for determining the connection pattern and the correct connection pattern for the connection endings connecting the Korean language phrases, By selecting the correct English conjunction for the role, it has the effect of generating the correct phrase-level English sentence.

Claims (6)

In the apparatus for determining connection information of English idioms applied to the Korean-English translation system, Morphological analysis means for performing morphological analysis of Korean sentences to be translated; Syntactic analysis means for performing structural analysis of a phrase unit using the analyzed information of the morpheme analysis means; Dependency relationship extraction means for receiving syntax structure information of a Korean sentence from the syntax analysis means and extracting a combination of all of the phrases connectable through a connection ending; Connection pattern storage means for storing a connection pattern of Korean idioms and basic connection patterns having the most frequent English translation information for each connection ending; Korean frame information storage means for determining a connection pattern to be applied to the Korean sentence to be translated in the connection pattern storage means; Connection pattern searching means for extracting a connection pattern list applicable to the dependency relationship combinations of the phrases extracted by the dependency relationship extracting means using the connection pattern storage means and the Korean language information storage means; Connection pattern set selection means for determining a most suitable connection pattern set that can provide all of the English connection information of connection terms of Korean sentences to be translated among the connection patterns searched by the connection pattern searching means; English generation information reflecting means for providing connection information and generation information of an English idiom corresponding to a Korean idiom using the connection information of the connection pattern set selected by the connection pattern set selecting means; Partial band sentence generation means for performing an English translation of a short sentence unit for each phrase received from the English generation information rebel means; Band sentence refining means for performing skipped element restoration work and style correction work such as sex, number, tense match information and subject for English sentence generation for each phrase translated into short sentences from the partial band sentence generation means; Morphological generation means for generating an English morpheme by receiving refined phrases from the band sentence refining means Device for determining the connection information of the English language phrases comprising a. In the connection information determination method of the English language phrases applied to the apparatus for determining the connection information of the English language phrases, Extracting combinations of all phrases using the syntax structure information of the Korean sentence; A second step of extracting a connection pattern list from a connection pattern database with respect to the connection relationship combination of the extracted phrases; A third step of determining whether a connection pattern extracted from the connection pattern list satisfies a Korean connection pattern condition; A fourth step of selecting a basic connection pattern corresponding to the extracted terminology from the connection pattern database if the Korean connection pattern condition is not satisfied as a result of the determination in the third step; A fifth step of selecting a corresponding connection pattern if the Korean connection pattern condition is satisfied as a result of the determination in the third step; A sixth step of determining whether all of the connection patterns to be applied to the connection relationship combinations of all extracted terms are selected; As a result of the determination in the sixth step, when all of the connection patterns to be applied to the connection relationship combination of all the phrases are selected, the connection pattern set selector for the connection relationship of the phrases of the Korean sentence to be translated from the list of selected connection patterns A seventh step of determining the most suitable set of connection patterns capable of providing all of the English access information; An eighth step of skipping to the second step if the connection pattern to be applied to the connection relationship combination of all the phrases is not selected as the result of the sixth step; And A ninth step of providing, in the English generation information reflector, using the connection information of the selected connection pattern set, access information and generation information of an English language phrase corresponding to a Korean language phrase How to determine the connection information of the English vocabulary, including. The method of claim 2, The first step is, A tenth step of performing morphological analysis and syntax analysis on the input Korean sentence; And And an eleventh step of extracting linkage combinations of linkable phrases using dependency information between Korean phrases from syntax structure information of a Korean sentence to be translated. The method of claim 2 or 3, The second step, A twelfth step of checking whether a suitable connection pattern exists in the connection pattern database with respect to the extracted connection phrase combination; And And a thirteenth step of extracting a connection pattern list by using the connection relation combination of the extracted phrases. The method of claim 2, The connection pattern database in the fourth step, A method for determining English phrases connection information comprising storing a connection pattern of Korean idioms and a basic connection pattern having the most frequent English translation information for each connection ending. In the English terminology connection information determination device having a microprocessor, A first function of extracting combinations of all phrases using syntax structure information of a Korean sentence; A second function of extracting a connection pattern list from a connection pattern database with respect to the connection relationship combination of the extracted terms; A third function of determining whether a connection pattern extracted from the connection pattern list satisfies a Korean connection pattern condition; A fourth function of selecting a basic connection pattern corresponding to the extracted terminology from the connection pattern database if it is determined that the third function does not satisfy the Korean connection pattern condition; A fifth function of selecting a corresponding connection pattern if the Korean connection pattern condition is satisfied as a result of the determination of the third function; A sixth function of determining whether all of the connection patterns to be applied are selected for the connection relationship combinations of all extracted terms; As a result of the determination of the sixth function, if all of the connection patterns to be applied to the connection relationship combination of all the phrases are selected, the connection pattern set selector for the connection relationship of the phrases of the Korean sentence to be translated from the list of selected connection patterns A seventh function of determining a set of most suitable connection patterns capable of providing all of the English connection information; An eighth function of skipping to the second function if the connection pattern to be applied to the connection relationship combination of all the phrases is not selected as a result of the determination of the sixth function; And A ninth function of providing connection information and generation information of an English language phrase corresponding to a Korean language phrase using the connection information of the selected connection pattern set in an English generation information reflector; A computer-readable recording medium having recorded thereon a program for realizing this.