KR20010103411A - Foreign-language translation system and method due to improving the translation efficiency - Google Patents

Abstract

The present invention relates to an English-Korean translation system and method with improved translation efficiency, and more particularly, to detect a terminating code candidate of a first language sentence from a first language document including a plurality of sentences, and to determine a termination portion. A sentence termination determiner for distinguishing sentences, an automatic word generator for recognizing and analyzing a word in the first language sentence and calculating a part-of-speech corresponding to each word and a second language word corresponding to the word, and the automatic word generator And a sentence structure generator that calculates a coupling relationship, sentence structure, and translation rate between words in the first language sentence.

The English-Korean translation system and method with improved translation efficiency according to the present invention display sentence component information and sentence structure information along with the translation of the first language sentence, and display the translation rate of the word in the sentence, the meaning of the word, and the sentence. It is possible to accurately grasp the structure and composition of the first language sentence, the completeness of the translation, and contribute to the improvement of the vocabulary of the user.

Description

Foreign-language translation system and method due to improving the translation efficiency

The present invention relates to an English-Korean translation system and method with improved translation efficiency by a linguistic based algorithm.

Generally, the English-Korean translation system consists of three stages: English interpretation, conversion, and Korean generation.

In particular, the English interpretation stage, which determines the performance of the English-Korean translation system, includes three processes: English stemmer, syntax parser, and semantic generator.

The English morpheme analyzer performs a function of designating an appropriate part-of-speech for each word constituting an English sentence, and the syntax analyzer receives a part-of-speech sequence as an input and analyzes a sentence structure.

Such English-Korean translation system, when analyzing long English sentences longer than 20 words, attempts to parse the whole sentence at once, not only takes a long analysis time, but also analyzes the structure due to problems such as syntax ambiguity processing. It is very likely to fail and result in incorrect translation.

In addition, in the conventional English-Korean translation system, the user cannot know the grammar information such as the sentence structure, the component, and the part-of-speech for the sentence in the English sentence to be translated. No information was known.

On the other hand, even if a user attempts to determine the meaning of a word in the sentence while translating the long English sentence, if the English dictionary does not exist in the surroundings, it causes inconvenience of not knowing the word.

In addition, since the English sentence is not completely translated when translated into Korean, the current situation is that the user does not fulfill the function as an English-English translation system by only being able to translate directly even if it takes time.

Accordingly, the present invention has been made in view of the above problems, and by using a new algorithm and data structure based on linguistic research, and using optimized grammar and dictionary, excellent throughput and fast processing for long sentences Translation speed can be achieved.

According to the present invention, it is possible to search for a word in a sentence and to facilitate the user's convenience.

The sentence structure information display function according to the present invention enables to grasp various information about the translation rate and grammar, thereby enabling more efficient English-Korean translation.

1 is a block diagram of a translation system according to a preferred embodiment of the present invention.

FIG. 2 is a detailed block diagram of a storage device using the English-Korean translation system of FIG. 1. FIG.

Figure 3 is a flow chart for explaining a method for translating English into Korean in the English-Korean translation system according to a preferred embodiment of the present invention.

4 is a flowchart illustrating a method for calculating sentence termination by an English-Korean translation system according to the present invention.

5 is a flowchart illustrating a method of calculating words by the English-Korean translation system of the present invention.

6 is a flowchart illustrating a method for calculating a sentence structure by the English-Korean translation system of the present invention.

Figure 7 is an exemplary screen display on the output device the results translated by the English-Korean translation system according to the present invention.

<Explanation of symbols for the main parts of the drawings>

100: input and output device 110: processor

120: memory 122: sentence termination grasp

124: Automatic Word Generator 126: Sentence Structure Generator

130: storage device 210: English word dictionary

220: English word collation dictionary 230: frequency dictionary

240: Pattern Dictionary 250: English Word Combined Dictionary

260: termination sign dictionary 270: termination pattern dictionary

According to a preferred embodiment according to an aspect of the present invention for implementing the present invention, the step of retrieving the sentence termination code of the input first language document and separating the sentences and sentences based on the sentence termination code, the division Calculating a part of speech and a second language corresponding to the first language by dividing the sentence by word, and identifying structure information of a sentence composed of a combination of the divided words. Can be provided.

According to a preferred embodiment provided with a language translation method, the step of distinguishing between the sentence and the sentence step of detecting a sentence termination code candidate of the input first language document using a termination code dictionary, the sentence termination using a termination pattern Calculating a likelihood of termination of a sign candidate, wherein the likelihood of termination is determined to be the end of a sentence when a termination level is equal to or greater than a predetermined level, and distinguishing between a sentence and a sentence based on the likelihood of termination; When there is a sentence termination code between the left and right marks, the name further includes excluding the sentence termination code candidate.

According to another preferred embodiment provided with a language translation method, dividing an input first language document by words using a morpheme analyzer and restoring the originality of the word, wherein the originality of the word is used by the mother language of the first language. Denoting the basic parts of speech of pronouns, nouns, verbs, adverbs, and adjectives,-extracting the separated words and determining the parts of speech using taggers, using a frequency dictionary, and using frequency of the input words Detecting and storing, wherein the frequency of use indicates a lexical level of a word quoted in a document input by a user, and compares the stored frequency with a frequency of words previously stored in a frequency dictionary. Calculating a word having a high word level based on the output; outputting the determined part-of-speech to a display device The method may further include generating a second language corresponding to the calculated word using a meaning generator.

According to yet another preferred embodiment provided with a language translation method, identifying a structure of a first language sentence using a pattern dictionary, wherein the pattern dictionary includes a coupling relationship between words; Determining completeness by calculating the applicability of each analysis rule from the structure of wherein the completeness is calculated by comparing a first language analyzed by the empirical rule stored in the storage device with a first language input by the user; Displaying the structure information of the sentence based on the completeness, calculating and displaying a translation rate based on the completeness, comparing analysis information of the input first language sentence structure with accumulated analysis information of the conventional sentence structure. Step by using the analysis rules used for sentence analysis, averaging the applicability of the displayed sentence structure It further includes the step of calculating.

According to a preferred embodiment of the present invention, a terminating candidate of a language sentence existing between the sentences is detected from an input language document, and a sentence terminating region is recognized from the sentence terminating candidate. Recognizes and analyzes words in each sentence from sentences separated from the sentence termination determiner for separating sentences, and a part-of-speech corresponding to each of the analyzed words and a second language word corresponding to the word. An automatic word generator for calculating, a sentence structure generator for calculating a combined relationship between words and words in the language sentence, a sentence structure and a translation rate based on the automatic word generator, a language word connected to the memory and a part of speech and a sentence of each word Storage device that retrieves and stores data information to identify structure and sentence termination In addition, the storage device determines a part of speech of a selected word and includes a word word collation storage unit and a language word dictionary that contain information on the suitability of the word based on the part of speech that is suitably entered before and after the word. A frequency storage unit configured to determine and store the frequency of the word based on the input word, and a termination pattern storage unit configured to store information regarding recognition of the termination code of the language sentence based on the termination code stored in the termination code storage unit. A language translation system can be provided.

According to a preferred embodiment according to another aspect of the present invention, detecting a sentence termination code candidate of the first language document using a termination code dictionary, calculating the possibility of termination of the sentence termination code candidate using a termination pattern Wherein the possibility of terminating is determined to be the end of a sentence when a level is greater than or equal to a predetermined level by assigning a level to the terminating code using a terminating dictionary, and distinguishing between the sentence and the sentence based on the possibility of terminating. A sentence separation method can be provided.

According to a preferred embodiment according to another aspect of the present invention, using the morpheme analyzer to separate the input language sentences by words and to restore the original form of the word, wherein the original form of the word is the mother of the first language Denoting the basic parts of speech of pronouns, nouns, verbs, adverbs, and adjectives used,-Detecting and storing the frequency of use of the input word using a frequency dictionary-The frequency of use is a document input by a user Indicating a vocabulary level of words recited within-comparing the stored frequency with words previously stored in the frequency dictionary, calculating a higher word level word based on the comparison, corresponding to the calculated word Generating a second language word, and outputting the calculated word and a second language corresponding to the word to a display device; How words can be provided which comprises.

According to a preferred embodiment according to another aspect of the invention, the step of identifying the structure of the first language sentence using a pattern dictionary, wherein the pattern dictionary includes a coupling relationship between words-, the identified Calculating the applicability of each analysis rule from the structure of the first language sentence to determine completeness, wherein the completeness compares the first language analyzed by the empirical rule stored in the storage device with the first language input by the user. And calculating the first language sentence structure into sentence component information and sentence structure information, outputting the sentence component information to a display device, and outputting structure information of the sentence to the display device based on the completeness. According to the present invention, a sentence structure display method comprising the step of calculating the translation rate based on the completeness and output to the display device can be provided. .

Hereinafter, with reference to the accompanying drawings will be described in detail a preferred embodiment of the present invention.

1 is a block diagram of an English-Korean translation system according to a preferred embodiment of the present invention.

Referring to FIG. 1, an English-Korean translation system according to the present invention is used in a computer system including an input / output device 100, a processor 110, and a memory 120, and the memory 120 includes a document termination identifier 122, An automatic word generator 124 and a document structure generator 126 are included.

Therefore, when an English document is inputted by the input / output device 100 by the user according to the present invention, the sentence terminator 122 provided in the memory 120 tracks a candidate for termination of the English document and matches the sentence. The sentences are distinguished and the translated results are presented to the user through the input / output device 100.

Here, the input English sentence is stored in the memory 120 through the bus 105, and the translated Korean sentence is output to the display device through the bus 105.

In addition, it is preferable that the document includes a sentence of reward.

Next, the automatic word generator 124 divides the words in the document to calculate a word suitable for the part-of-speech and interpretation of each word, and displays Korean of the word.

Next, the document structure generator 126 analyzes sentences using rules regarding the combination of words and words in the English sentence to calculate the completeness of the sentences, arranges each word according to Korean grammar, and translates the sentence. Judge.

The input / output device 100 includes an input device unit including a keyboard or a mouse for inputting an English sentence to be translated, and an output device unit including a CRT monitor or a display for displaying the translated result.

In addition, the memory 120 is connected to the storage device 140 that stores information about words and parts of speech, document structure, and document termination, and exchanges data through data bus lines. Segmentation based on parts of speech, sentence structure, and sentence ending can be calculated.

The input / output device outputs a Korean sentence converted by the user through an input device through the sentence termination detector 122, the automatic word generator 124, and the sentence structure generator 126. To print.

Here, the sentence termination determiner 122, the automatic word generator 124 and the sentence structure generator 126 constituting the memory 120 are the greatest features of the present invention.

The sentence terminator 122 divides the document input by the user into a sentence and a sentence, finds the end of the sentence, and separates it from the next sentence.

The automatic word generator 124 separates the sentence for each word and calculates a word suitable for the user's level of the part of speech and the level of use of the word. In addition, the Korean language corresponding to each word is calculated and displayed on the input / output device 100.

The sentence structure generator 126 calculates a sentence component and a sentence structure based on the words generated from the automatic word generator 124.

FIG. 2 is a detailed block diagram of a storage device by the English-Korean translation system of FIG. 1.

The storage device 140 is connected to the memory 120 of the English-Korean translation system by a data bus line, and the sentence terminator 122, the automatic word generator 124, and the sentence structure generator 126 of the memory 120 are provided. Processes and transmits and receives various data files stored in advance in the storage device 140.

The storage device 140 is a word storage unit 210 for storing data for English words, a word collation storage unit 220 for storing data about arranging a plurality of words at once, the frequency of use of the input word It includes a frequency storage unit 230 for storing the data for.

In addition, the storage device 140 may include a pattern storage unit 240 for storing data about usage forms of each word, a word combining storage unit 250 for storing data about the degree of combining English words, and a terminating code. A terminating code storage unit 260 for storing data and a terminating pattern storage unit 270 for storing data for terminating recognition of an English sentence.

The storage units input information necessary for realizing the present invention to the storage device 140 as a data file, and when the English translation is required, the data stored in each storage unit performs a predetermined operation in the memory. Translate English sentences into Korean sentences.

The word storage 210 stores a word corresponding to an English word in an English sentence in advance, and is defined as a Korean word corresponding to a new English word by a user.

The word colocation storage unit 220 stores data related to arranging a plurality of words at once and prioritizing them.

In the frequency storage unit 230, the degree of difficulty for each word of the used English sentence is set in advance and stored as a data file.

Here, the "frequency" refers to the lexical level (difficulty level) of the word, and the frequency dictionary stores the frequency of each word in advance.

For example, in the English-Korean dictionary in circulation, words such as 'go', 'take', and 'school' are indicated by a (??) in front of the word and refer to a low difficulty word.

Also, words such as 'casual', 'fold', and 'sanitary' are marked with (ː) to indicate a slightly higher difficulty, while 'confound' becomes (*) and 'inveterate' becomes (), etc. It refers to more difficult words.

Therefore, the frequency of each word is displayed and stored in the frequency dictionary in advance.

Therefore, when each word is recognized in the input English sentence, the corresponding frequency is checked and a higher level word is selected, which can be selected / deselected by the user's setting.

The pattern storage unit 240 sets and stores the degree of coupling of each language unit in advance, and the ending pattern storage unit 270 sets and stores a rule regarding the ending in order to find the ending of the sentence.

Here, the term 'pattern' is to set the relationship between language units so that it is possible to determine whether the units of each language used in the language can be combined with each other.

Accordingly, the pattern storage unit 240 may be used to analyze the suitability of the word as well as the relationship between the word and the word.

The word combining storage unit 250 has a concept similar to that of the pattern storage unit 240, and sets and stores rules related thereto so that words and words are properly combined.

For example, if the verb 'go' is followed by the verb 'take', this is considered a grammatical error.

In addition, when verbs are followed by words that can be used as verbs or nouns, rules about the adequacy of the association between these words are set to infer the correct part-of-speech of the available words of the following verbs or nouns.

The termination code storage unit 260 data-stores the symbol representing the termination of the sentence in advance.

For example, '!', '?', '.' To symbols such as

In addition, the storage device stores the English sentence analysis rule information for storing the information on the English morpheme combining / analysis rule information for storing the information about the degree of association and analysis of the morphemes of the English words in advance, and the information about the analysis rule of the English sentence used It further comprises means.

For example, the English morpheme combining / analysis rule information may be stored in the storage device by presetting a morpheme sequence combining a word and a part of speech corresponding to the word, and storing the sentence termination detector 122 and the automatic word generator 124. ), A part-of-speech word is output to the user output device.

For example, the English sentence analysis rule storage means is based on the morphemes for the words stored in the English morpheme combining / analysis rule storage means, by setting in advance a rule for analyzing whether the structure is suitable for the structure as a sentence by Store in the storage device.

Therefore, when the English sentence input by the user is arranged in the morpheme sequence by the English morpheme combining / analysis rule storage means, it is analyzed whether the arrangement of the morpheme sequence is suitable as the structure of the sentence.

Hereinafter, a description will be given of a translation process for generating English in Korean according to the present invention.

3 is a flowchart illustrating a method of translating English into Korean in an English-Korean translation system according to a preferred embodiment of the present invention.

First, when a user inputs a document to be translated (step 310), a sentence termination code candidate is detected between a sentence and a sentence by a sentence termination determiner and recognizes the end of the sentence based on the sentence termination code candidate to recognize the sentence between sentences. (Step 320).

Next, based on the divided sentences, the sentence is divided into words to calculate a part-of-speech and an optimal word of each word (step 330).

Next, the sentence structure information of the sentence composed of the words and the translation rate of the translated sentence are calculated based on the calculated word (step 340).

Next, the translated result and the additional information are output together on the screen (step 350).

Here, the additional information relates to parts of speech, sentence structure information, sentence component information, and translation rate of each word.

4 to 6 will be described in more detail.

4 is a flowchart illustrating a method of calculating sentence termination by an English-Korean translation system according to the present invention.

First, when a user inputs a document by the document input step (step 310) of FIG. 3, a sentence terminating candidate is searched for to distinguish a plurality of sentences included in the input document (step 410), and the sentence termination is performed. The possibility of termination is calculated based on the sign candidates (step 420).

Next, it is determined whether the sentence is terminated based on the calculated sentence termination possibility, and when the predetermined degree or more is reached, the sentence is terminated and the sentence is separated from the sentence (step 430). Analyze the words.

In this case, a punctuation dictionary in which data regarding sentence termination is stored using preset data is used, which is based on a sentence termination partial search.

Therefore, the ending portion of the previous sentence and the ending portion of the current sentence are searched and extracted from one word following the previous sentence to a word including the ending portion of the current sentence as one sentence.

The sentence termination portion search is performed as follows.

First, the sentence termination part is searched among documents input by the user. At this time, the sentence termination code dictionary is retrieved to determine whether the sentence termination dictionary is consistent with the contents of the sentence termination dictionary.

As described above, the sentence terminator is '!', '?', '.' Codes that signify the end of a sentence, such as the data is stored and stored, and the end of the sentence is searched for by comparing whether such a code exists in the document.

Secondly, the characteristics of the sentence termination code part itself are searched and set as the sentence termination code candidate.

Here, the characteristic of the sentence terminator itself includes a case where a space after the sentence terminator is present without being used alone.

Third, the left and right characteristics of the sentence with the termination code are examined by using the termination pattern storage unit among sentence termination code candidates, and finally recognized as the sentence termination part.

For example, in John F. Kennedy's case, F is followed by a '.' Although a sentence termination code candidate exists, it is qualified as a sentence termination code candidate, but when the left and right seals and names appear the same, the initial portion of the name is displayed. Therefore, the sentence termination portion is excluded.

Fourthly, the above process is carried out only for the special and complex use of the sentence termination code.

For example, '.' Is also the most complicated because it also appears between names and names, and '!' Or '?' Etc., since the recognition is clear at the end of the sentence, it is simple and can be recognized as the end of the sentence only through the first and second steps.

5 is a flowchart illustrating a method of calculating words by the English-Korean translation system of the present invention.

First, when a sentence and a sentence are distinguished from each other in FIG. 4, the sentence termination determiner 122 in the memory retrieves each word used in one sentence using the English word dictionary 210 (step 510).

Next, the part of speech of each word is determined using the pattern dictionary 240 and the English word collation dictionary 220 (step 520).

In this case, the pattern dictionary 240 analyzes the suitability of the combination of words and words, the suitability of the word in the sentence, the part of speech and the suitability of the meaning of the word, and the English word collation dictionary 220 Arrange multiple words at once and process them according to the rules of priority.

In addition, when extracting the part-of-speech, the part-of-speech information corresponding to the word is input to the storage device 140 in advance using a tagger, and the part-of-speech of the word is estimated by comparing with the word in the used sentence.

Next, the frequency of the word is recorded in the frequency dictionary 230, that is, the vocabulary level of the word is stored at the same time (step 530), while the sentence terminator 122 determines the vocabulary level of the word and a predetermined vocabulary. Compare levels (step 540).

If the vocabulary level of the word is higher, the word is calculated (step 550); otherwise, the predetermined word is calculated (step 560).

That is, the words in the English sentence is stored in the frequency dictionary 230, the frequency of the word is stored, and compared with the frequency for the word specified in advance.

As a result of the comparison of the frequencies, if the frequency of the words in the English sentence is higher than the frequency of the predetermined word, the words in the English sentence are generated as it is.

In this case, it is preferable that the words in the English sentence and the word specified in advance include the same meaning when translated into Korean.

For example, 'reveal' is used in Korean to mean 'disclose', and similar words include 'disclose', 'divulge' and 'betray'.

Looking at the difficulty of the words in the English-Korean dictionary, 'reveal' is indicated by (ː), 'disclose' by (ː), 'divulge' by (), and 'betray' by (). Here, it can be seen that 'reveal' and 'disclosure' are of similar difficulty and lower than 'betray' and 'divulge'.

Therefore, based on the above fact, the storage device has 'reveal' as 'step 1', 'disclose' as 'step 2', 'divulge' as 'step 3' and 'betray' as 'step 4' When it is marked and stored in advance, if the word in the English sentence is 'disclose', 'divulge' is generated. If the word is 'divulge', 'betray' is generated.

6 is a flowchart illustrating a method of calculating a sentence structure by the English-Korean translation system of the present invention.

When the words are calculated in FIG. 5, it is necessary to analyze a sentence which is a combination of the following words.

First, when a word is calculated by FIG. 5, the automatic word generator 124 retrieves the part-of-speech information of the word of the English sentence (step 610).

At this time, to retrieve the part-of-speech information of a word, an English word dictionary 210, a pattern dictionary 240, and an English word collation dictionary 220 are used.

Here, the descriptions of the English word dictionary 210 and the pattern dictionary 240 are as described above, and the English word collation dictionary 220 includes information on arranging a plurality of English words in a line. .

Therefore, when the association relationship between words and words is identified by the pattern dictionary 240, the words are arranged according to the structure of the sentence, and the English word collation dictionary 220 rearranges the words in the order of words. It plays a role.

Next, the structure of the English sentence is determined using the English word combining information dictionary 250 based on the parts of speech of each extracted word (step 620).

In the step of identifying the sentence structure (step 620), the structure of one sentence is identified, and the role of a word in one sentence is determined in the sentence structure.

That is, the information on the combination of the words in the English word combination information dictionary 250 is retrieved in this step (620) to determine the structure of the sentence.

In the sentence structure information analyzed in step 620, the sentence structure analysis information is divided into sentence component information for dividing an English sentence into component elements and sentence structure information for dividing the English sentence into phrases or clauses and then output on the screen.

The sentence component information divides a sentence into a subject and a verb, and the sentence structure information divides the English sentence into parentheses based on the sentence component information. The sentence component information is divided into morphemes by parts of speech extracted in the sentence. Understand the structure of sentences based on the words placed.

Next, the completeness and the translation rate of the sentence are determined from the steps 610 and 620 (step 630).

The completeness of the present invention is to calculate the empirical rule and the data analyzed in advance and compare the information and the information of the input English sentence.

In addition, by continuously using the English-Korean translation system according to the present invention, the analysis data for the sentence structure is continuously input to the storage device 140 to become important basic data for determining the completeness of the sentence structure to be used later. The completeness is judged more accurately.

In addition, the translation rate is to determine the accuracy of the translated sentence, the criteria for calculating the translation rate in the present invention is as follows.

First, the translation rate is based on translation.

Accordingly, the user may determine how much difference occurs by comparing the translated result with the literal sentence stored in the storage device 140, which is divided into semantic accuracy and grammatical accuracy.

The semantic accuracy refers to how accurately the meaning of the original text is transmitted to the user and means semantic criteria in linguistics.

Therefore, the semantic accuracy refers to the syntactic criterion in linguistics as the grammatical accuracy refers to how accurately the structure of the original text is analyzed grammatically and properly corresponded to the structure of the translated language.

In the present invention, in order to calculate the translation rate, the semantic accuracy and grammatical accuracy of the translated sentences are evaluated by distinguishing the semantic accuracy and grammatical accuracy in the following six steps, and scores are given for each step, and the results calculated by summing Outputs

That is, the following six steps are previously stored in the storage device, and the translation rate calculated by the processor is compared with the level stored in the storage device, and the corresponding level is output to the user's display device.

Step 1: 0 points-Translations without any information related to the original.

Step 2: 1 point-translations that provide very little information related to the original text.

Step 3: 2 Points-Translation to get some information about the text.

Level 4: 3 points-Translations that can understand more than half of the text.

Step 5: Four Points-A translation that has some errors but generally understands the meaning of the original text.

Step 6: 5 points-Translations with the same or equivalent comprehension as literal texts.

Here, step 1 is not translated at all, step 2 is only capable of translating words, step 3 is able to grasp phrases somewhat, step 4 is somewhat translated from step 4, and final step 6 is almost original Results in a perfect translation that matches.

Thus, steps 5 and 6 are successful translations and the rest are considered translation failures.

7 is an exemplary view of a screen displaying a result translated by the English-Korean translation system according to the present invention on an output device.

As shown in FIG. 7, a title bar, a menu bar, and a tool bar are provided on the screen from the top. An English document input window is located on the right side of the toolbar, and a word information window is on the left side. The "Translation Result Window" is installed below.

The menu bar includes files, edits, views, translations, tools, windows, and help.

The translation menu may include a translation of a document after a current caret position, a translation of a current caret position sentence, a selection region translation, and a translation of the translated document in a new window.

The translation of the document after the current caret position is used to execute the translation of the current document and to translate all the documents after the location of the current caret, and the translation of the sentence after the current caret position is used to translate the next sentence with the current caret. do.

The translation of the selection region designates a selection region with a mouse to execute the translation of the portion, and moves the already translated sentence to a new window in the translation window document new window, and the contents of the translation window are deleted.

The tool menu largely includes a user dictionary and an option, and the user dictionary sets and sets words which are not stored in the English word dictionary and word meanings corresponding thereto.

The option includes a translation output format and a dictionary option. The dictionary option may be set by the user to automatically set the frequency of automatic word search frequency, find one by one, and find synonyms and thesaurus in the dictionary information window option.

The "tool bar" includes a current document translation tool, a translation window switching tool, a word information window switching tool, and the like.

Therefore, when a document is to be translated, a translation is immediately performed when the mouse is positioned at the beginning of the document to be translated by the user and the current document translation tool is clicked.

In addition, as the translation window switching tool and the word information window switching tool are pressed with a mouse, the English document input window may be used as all screens, and the word information window and the translation window may be divided and used as shown in FIG. 6.

When a user inputs an English document into an English document input window and presses the current document translation tool with a mouse, the translated result is displayed in a translation window at the bottom.

In this case, as shown in FIG. 6, the translation window displays a original sentence, a sentence in which the original sentence is translated into Korean, a sentence component, a sentence structure, and a part of speech of each word.

The top of the translation window represents an English sentence to be interpreted, the Korean translated below the English sentence is displayed, the sentence component, sentence structure, parts of each word are displayed in turn.

In addition, a translation indicator, which is an indication of the degree of the translated result, is calculated using a bar-shaped indicator on the left side of the translation window. That is, if it is displayed up to the upper part of the indicator, the translation is good, and if it is displayed below, it is determined that the translation has failed.

The word information window on the left side of the screen is to retrieve and display the words stored in the English word dictionary, the user writes a word to know and press the word information icon to display the information about the word.

The most important function of the word information window is that the word used in the sentence of the English document input window is automatically calculated and displayed on the word information window.

That is, when a sentence is set in the written document of the English document input window with a mouse, information about each word used in the sentence is automatically displayed in the word information window.

Accordingly, the user can more accurately understand the information about the words in the sentence, thereby improving the learning efficiency.

In the above, the process of converting an English document into a Korean document was described with reference to a preferred embodiment of the present invention. However, the present invention may also convert between different languages.

In addition, it will be understood by those skilled in the art that the present invention may be variously modified and changed without departing from the spirit and scope of the present invention as set forth in the claims below.

The English-Korean translation system and method with the improved translation efficiency as described above automatically calculates the words used in the sentences and displays the word meanings corresponding to the words, so that the user can clearly understand the meaning of each word, thereby having a great learning effect. .

Further, according to the present invention, the completeness of the sentence is more clearly defined, and the translation rate is displayed on the screen, so that the efficient translation and the user can directly check the degree of translation, thereby determining the reliability of the translation result.

In addition, according to the present invention, it is possible to search and confirm a word that the user wants to know from the word information window of the screen, and to view the analysis information and the translation result according to the selection of the analysis information and the translation result of the translation window, Part-of-speech, sentence components, and sentence structure of words within a letter are also displayed at the same time, so that sentences can be understood more clearly.

Claims (34)

A language translation method comprising a memory having a program and a processor connected to the memory to translate a first language into a second language, Retrieving a sentence terminator of an input first language document and distinguishing a sentence from a sentence based on the sentence terminator; Calculating the part-of-speech of the word and a second language corresponding to the first language by dividing the divided sentence by each word; Identifying structure information of a sentence composed of a combination of the divided words Language translation method comprising a. The method of claim 1, Terminator with '!', '?', And '.' The language translation method characterized by the above-mentioned. The method of claim 1, Distinguishing the sentence from the sentence Detecting a sentence termination code candidate of the input first language document using a termination code dictionary; Calculating a likelihood of termination of the sentence termination code candidate using a termination pattern; Distinguishing between sentences based on the possibility of closing Language translation method comprising more. The method of claim 3, The possibility of termination And determining the end of the sentence when the level is greater than or equal to the predetermined level using the ending code level. The method of claim 3, Excluding sentence termination code candidates when there is a sentence termination code between left and right characters and names Language translation method comprising more. The method of claim 1, Dividing the input first language document by words using a morpheme analyzer and restoring the original form of the word; Extracting the separated word and determining a part-of-speech of the word using a tagger Language translation method comprising more. The method of claim 6, The circle of the word is A method of translating a language characterized by representing pronouns, nouns, verbs, adverbs, and adjective root parts of speech used in the first language. The method of claim 6, Detecting and storing a frequency of use of the input word using a frequency dictionary; Comparing the stored frequency with a frequency of words previously stored in a frequency dictionary; Calculating a word having a high word level based on the comparison Language translation method comprising more. The method of claim 8, The frequency of use And a lexical level of the words quoted in the document input by the user. The method of claim 6, Outputting the determined part-of-speech to a display device Language translation method comprising more. The method of claim 6, Generating a second language corresponding to the calculated word using a meaning generator Language translation method comprising more. The method of claim 1, Identifying a structure of a first language sentence using a pattern dictionary; Determining completeness by calculating applicability of each analysis rule from the identified language sentence structure; Displaying structure information of a sentence based on the completeness Language translation method comprising more. The method of claim 12, Calculating and displaying a translation rate based on the completeness Language translation method comprising more. The method of claim 12, The pattern dictionary is A language translation method comprising a binding relationship between words. The method of claim 12, The completeness And a first language analyzed by the empirical rules stored in the storage device is calculated by comparing the first language input by the user. The method of claim 12, Comparing the analysis information of the input first language sentence structure with the accumulated analysis information of the conventional sentence structure; Language translation method comprising more. The method of claim 12, The language sentence structure is Sentence component information divided into sentence component elements; Sentence structure information for dividing the sentence into parentheses based on the sentence component information Language translation method consisting of. The method of claim 17, Displaying the sentence component information and sentence structure information on a screen; Language translation method comprising a more. The method of claim 12, Calculating completeness by averaging the applicability of the segmented sentence structure using the analysis rule used for sentence analysis Language translation method comprising more. A language translation system comprising a memory having a program and a processor connected to the memory to translate a first language into a second language, A sentence termination determiner for detecting a termination code candidate of a language sentence existing between the sentences among input language documents, and recognizing a sentence termination region from the sentence termination code candidate to distinguish a sentence from a sentence; An automatic word generator for recognizing and analyzing a word in each sentence from sentences separated from the sentence terminator, and calculating a part-of-speech corresponding to each analyzed word and a second language word corresponding to the word; A sentence structure generator that calculates a combined relationship, sentence structure, and translation rate between words in a language sentence based on the automatic word generator. Language translation system. The method of claim 20, A storage device connected to the memory for retrieving and storing data information to grasp language words and parts of speech, sentence structure, and sentence termination parts. Language translation system made further comprising. The method of claim 21, The storage device A language word colocation storage unit that determines the part-of-speech of a selected word and contains information on the suitability of the word based on the part-of-speech of the word to be suitably entered before and after the word. Language translation system comprising a. The method of claim 22, Frequency storage unit for determining and storing the frequency of the word based on the words entered in the language word dictionary Language translation system made further comprising. The method of claim 22, Termination pattern storage unit for storing information on the termination code recognition of the language sentence based on the termination code stored in the termination code storage unit Language translation system made further comprising. In the sentence separation method of distinguishing one sentence from another sentence based on a first language document including a plurality of sentences input by a user for translation based on a language translation method, Detecting a sentence terminating candidate of the first language document using a terminating dictionary; Calculating a likelihood of termination of the sentence termination code candidate using a termination pattern; Distinguishing between sentences based on the possibility of closing Sentence separation method comprising a. The method of claim 25, The possibility of termination A sentence separation method, characterized in that it is determined that the sentence is terminated when the termination code is given a level by using the termination code dictionary and is above a predetermined level. In the word display method for outputting a specific word in the first language sentence based on the language translation method to the word information window of the user's display device, Dividing the input language sentences by words using a morpheme analyzer and recovering the prototype of the words; Detecting and storing a frequency of use of the input word using a frequency dictionary; Comparing the stored frequency with a word previously stored in the frequency dictionary; Calculating words above a certain vocabulary level based on the comparison; Generating a second language word corresponding to the calculated word; Outputting the calculated word and a second language corresponding to the word to a display device; Word display method comprising a. The method of claim 27, The circle of the word is A word display method, characterized in that it represents a fundamental part-of-speech of pronouns, nouns, verbs, adverbs, and adjectives. The method of claim 27, The frequency of use And a lexical level of the words quoted in the document input by the user. A sentence structure display method for analyzing and displaying a structure of a first language sentence based on a language translation method, Identifying a structure of a first language sentence using a pattern dictionary; Determining completeness by calculating an applicability of each analysis rule from the identified first language sentence structure; Dividing the first language sentence structure into sentence component information and sentence structure information; Outputting the sentence component information to a display device Sentence structure display method comprising a. The method of claim 30, Outputting structure information of a sentence to a display device based on the completeness Sentence structure display method comprising more. The method of claim 30, The translation rate is calculated based on the completeness and output to the display device The sentence structure display method further comprising. The method of claim 30, The pattern dictionary is Associative relationships between words Sentence structure display method comprising a. The method of claim 30, The completeness And a first language analyzed by the empirical rules stored in the storage device is calculated by comparing the first language input by the user.