KR20210126240A - System for learning structures of foreign languages - Google Patents

Abstract

The present invention provides a system for learning a foreign language structure, which can learn the foreign language sentence structure by chatting using messenger. The system includes a control module, a data collection module, a database, an image combination module, and a sentence analysis module. The control module may collect data about a sentence, provide data stored in a database to each module, or control an operation of each module. The data collection module may collect data on words constituting the sentence and provide it to the control module. The sentence analysis module may analyze a structure of a sentence by using artificial intelligence (AI) such as machine learning (sentence parsing).

Description

Foreign language sentence structure learning system {SYSTEM FOR LEARNING STRUCTURES OF FOREIGN LANGUAGES}

The present invention relates to a system capable of learning foreign language sentence structures. More particularly, it relates to a foreign language sentence structure learning system capable of learning a foreign language sentence structure through a chatting method using a messenger.

The arrangement positions of sentence components are different for each language. Therefore, in order to study a foreign language, it is necessary to learn how sentence components are arranged rather than simply memorizing sentences.

However, in the conventional foreign language learning system, only memorizing sentences or providing contents in a manner of listing words.

Therefore, research on a system that can maximize learning efficiency by learning the sentence structure of a foreign language rather than simply learning foreign language sentences is being conducted.

Republic of Korea Patent Registration No. 10-1906561 Republic of Korea Patent Registration No. 10-1853892

An object of the present invention is to provide a system capable of learning foreign language sentence structures.

The present invention provides a foreign language sentence structure learning system capable of learning a foreign language sentence structure through a chatting method using a messenger.

According to the present invention, it is possible to learn a word suitable for one's level, and it is possible to learn an English-style speech structuring method using the mother tongue.

1 to 15 are conceptual diagrams for explaining a system according to an embodiment of the present invention.

Since the present invention can have various changes and can have various embodiments, specific embodiments are illustrated in the drawings and will be described in detail in the detailed description.

However, this is not intended to limit the present invention to a specific embodiment, it should be understood to include all modifications, equivalents and substitutes included in the spirit and scope of the present invention.

Hereinafter, the present invention will be described in more detail with reference to the drawings. In the present specification, the same and similar reference numerals are assigned to the same and similar components in different embodiments, and the description is replaced with the first description. As used herein, the singular expression includes the plural expression unless the context clearly dictates otherwise.

In addition, the suffix "module" or "part" for the components used in the following description is given or mixed in consideration of only the ease of writing the specification, and does not have a meaning or role distinct from each other by itself.

Learners may be provided with language learning content through the user terminal. User terminals include mobile phones, smart phones, laptop computers, digital broadcast terminals, personal digital assistants (PDA), portable multimedia players (PMPs), navigation systems, slate PCs, and tablet PCs. PC), ultrabooks, wearable devices, for example, watch-type terminals (smartwatch), glass-type terminals (smart glass), mobile terminals such as HMD (head mounted display), digital TV, desktop A fixed terminal such as a computer, digital signage, and the like may be included.

The system 100 may include a control module 110 , a data collection module 120 , a database 130 , an image combination module 140 , a sentence analysis module 150 , and the like. Hereinafter, each module is separately described for convenience of description, but it is possible that the actual control module 110 is formed to include functions of other modules.

The control module 110 may collect data about a sentence, provide data stored in a database to each module, or control the operation of each module. To increase learning efficiency, the control module 110 may include a machine learning module. For example, it is possible to find the optimal image by collecting information such as the box size and color preferred by learners in the process of imaging sentence structures through the image combination module.

When exposing the sentence structure image, the control module 110 may provide a voice corresponding to the image together.

The data collection module 120 may collect data on words constituting a sentence and provide it to the control module 110 . For example, the data collection module 120 may collect voice data corresponding to words and store them in a database.

The database 130 stores various data for driving the system 100 . For example, the database 130 may store image data corresponding to words and a plurality of voice data corresponding to the words.

In addition, data on the color, size, and position of the image box may be stored in the database 130 . Such information can be set by the user and, in some cases, may be information obtained through machine learning.

The image combination module 140 matches an image box corresponding to each word (or group) and defines the size, color, position, and the like of the image box.

In addition, the image combination module 140 may define the position of the image box (group) of words in each set, and then define the image order and position between groups.

The sentence analysis module 150 may analyze the structure of a sentence by using artificial intelligence (AI) such as machine learning (sentence parsing).

According to an embodiment of the present invention, the system may further include a bullish pattern analysis module (not shown).

The stress pattern analysis module can determine the stress pattern of the sentence the learner wants to listen to by analyzing the voice output data of the sentence and analyzing the position where the speaker cuts the sentence, the output dB value, the tone or intonation, etc.

The control module may adjust the display interval or display time of boxes provided to the learner according to the pattern or sentence analysis result analyzed by the stress pattern analysis module.

The image structuring method of the English sentence structure may include a pre-data processing step (S100), a data processing step (S200), a sentence structure imaging step (S300), and a learning data provision step (S400).

The system of the present invention extracts main words by using the POS value and DEP value of the English sentence, sets a set for each word group, and uses the organized data to structure the English sentence image.

The sentence analysis module detects data on a POS (Part of speech) value of a sentence, a DEP (Dependency) value, and a word corresponding to the head of each word.

The POS value represents the sentence component value of each word.

POS values include ADJ(adjective), ADP(adposition), ADV(adverb), AUX(auxiliary), CONJ(conjunction), CCONJ(coordinating conjunction), DET(determiner), INTJ(interjection), NOUN(noun), NUM (numeral), PART(particle), PRON(pronoun), PROPN(proper noun), PUNCT(punctuation), SCONJ(subordinating conjuction), SYM(symbol), VERB(verb), X(other) SPACE(space), etc. may include.

The DEP value is a value indicating a relationship between words, and each word may be connected with an arrow having a starting point (tail) and an ending point (head) as shown in FIG. 4 .

In this embodiment, arrows connecting each word are used to allocate the DEP value, but this is only for explaining the relationship between the words. That is, even if the system does not indicate the relationship between words with arrows, if a meaning is connected from one word to the next or a sentence structure is connected, it is possible to set the side of the preceding word as the tail point and the side of the following word as the head point. do.

DEP values include acl (adjectival clause), acomp (adjectival complement), advcl (adverbial clause modifier), advmod (adverbial modifier), amod (adjectival modifier), appos (appositional modifier), ccomp (clausal complement), dobj (direct object). ), nsubj (nominal subject), det (determiner), relcl (relative clause modifier), and the like.

A child refers to a component located at the head point of an arrow connecting two components, and a head refers to a component at the starting point of an arrow connecting two components.

An ancestor refers to a component of a position from the head back to the head.

Descendant refers to a component of a position from child to child.

The control module 110 may process data in response to information such as the POS value, the DEP value, the child, descendant, and encestor described above, and set the arrangement, shape, color, size, etc. of the image box.

The data processing step (S200) includes a step (S210) of selecting main words from an English sentence, a step (S220) of specifying a set of words and tying the words of the same set into one box (S220), setting a box group and a box position It may include a step (S230) and the like.

In the step of selecting the main words ( S210 ), the main word is selected from the sentence by using the POS value, the DEP value of the above-described words, and the connection structure information between the components.

< Identification of the main word (is main) >

If it corresponds to the main word, the main word data value (is main) is set to true, and if it is not the main word, it is set to false.

The control module 110 has a DEP value of the words acl, acomp, advcl, advmod_v, advmod_v_up, appos, attr, aux_t, auxpass, cc_t, cc_c, ccomp, conj_t, cop, csubj, csubjpass, dative, dep, dobj, expl , intj, meta, nsubj, nsubjpass, oprd, obj, obl, parataxis, pcomp, pobj, preconj, punct, relcl, root, xcomp, select the corresponding word as the main word.

When the DEP value is conj, if the POS of the word corresponding to the head of the corresponding word is verb, the DEP value of the corresponding word is changed to conj_t (conj corresponding to the main word) and this is selected as the main word. When the DEP value is conj, if the POS of the word corresponding to the head of the corresponding word is not a verb, the DEP value is changed to conj_f (conj not corresponding to the main word) and it is determined as not the main word.

Here, the DEP values conj_t and conj_f are indicated for convenience of explanation, and the scope of rights is not limited to these terms.

In addition, when the DEP value is cc and the POS value of the word corresponding to the head of the word is verb, the control module 110 changes the DEP value to cc_c when the root is head and the root index is greater than its own index, and the root is head and if the root index is smaller than its own index, the DEP value is changed to cc_t. If the POS value of the word corresponding to the head is not a verb, the DEP value is changed to cc_f (cc not corresponding to the main word) and it is determined as not the main word.

Similarly, the DEP values cc_t and cc_f are indicated for convenience of explanation, and the scope of rights is not limited to these terms.

In addition, in the case of a word having a DEP value of prep or agent, when the DEP value of a word corresponding to a child of the word is pcomp, the control module 110 does not select a word having a DEP value of prep or agent as the main word. .

The control module 110 changes the DEP value to cc_c when the DEP value is npadvmod, advmod and the POS of the head is verb, when the root is the head and the root index is greater than the own index, and the root is not the head and the root index is the own index If it is less than, change the DEP value to advmod_v. Through this transformation, it is possible to handle the case where the adverb is separate from the adverb of the adverb.

If the DEP value is mark and the POS of the head is verb, if the head index is larger than the self index, the DEP value is changed to cc_c. Through this, it is possible to handle the case where a word (mark) such as as is in front of the adorning word.

In addition, the control module 110 changes the DEP value of the word corresponding to the head to prep or the DEP value of the word corresponding to the head to pobj_t (pobj corresponding to the main word) while the DEP value is pobj, and selects it as the main word. If the DEP value is pobj and the DEP value of the word corresponding to the head is not prep or agent, the DEP value is changed to pobj_f (pobj not corresponding to the main word), and it is determined as not the main word.

Also, if the DEP value is aux, the control module 110 changes the DEP value to aux_t if the lemma is have or be.

In addition, when the POS value of the head of the word 'of' is NOUN or PRON, the control module 110 excludes the word having the corresponding 'of' as the head from the main word.

<Check the depth of the main word>

If the head of 'of' is A, and the word having 'of' as head is B, the head of the word having B as head is changed to A.

The control module designates the depth of the word whose DEP value is the root as 0, and when the number of main words encountered when each word goes back to the root on the head is n, the depth of the corresponding word is designated as n. In this case, the word itself is not counted as the main word.

According to an embodiment of the present invention, the control module 110 may set the size and font of characters according to the depth value or the level value described below.

<Setting part value (part_main) and set value of words>

Determines which main word the words are part of.

The control module 110 first determines the index value as a part value in the case of the main word (true).

If the word does not correspond to the main word (false), the part value (part_main) of the word is determined by the following method.

If the word corresponding to the head of the word is the main word (true), the index value of the head is determined as the part value (part main). If the word corresponding to the head is not the main word (false), it is checked whether the word corresponding to the head of the head is the main word, and the index value of the corresponding main word is determined as the part value.

In the case of a word that is not a main word and whose DEP value is prep or agent, if there is a main word among the words having the word as head, the index value of the corresponding main word is determined as the part value. For example, in the case of 'in' with an index value of 12, the DEP value is prep while not the main word. Since the word 'home' having an index value of 15 among words having 'in' as the head corresponds to the main word, the index value 15 of 'home' becomes the part value of 'in'.

In the case of a word that is not a main word and whose DEP value is prep or agent, if there is no main word among the words that have the word as head, the index value of head is determined as the part value. For example, in the case of 'of' with an index value of 30, the DEP value is prep while not the main word. However, since the main word does not exist among the words having 'of' as the head (words with a head index value of 30), the head index value of 'of', 29, becomes the part value.

For easy understanding, corresponding words are indicated in red letters in FIG. 6 .

Continue to determine the part value in the same manner as above until the part value of all words is determined.

After determining the part values of the words, a set of words is set by using the main word and the determined part values.

The control module 110 sets the index value of the main word as a set value when it corresponds to the main word, and sets the part value (part_main) of the corresponding word as a set value when it is not the main word.

<Calculation of word level>

Set the level of words to determine the size, color, etc. of boxes or groups.

The level calculation of words is determined according to the following rules in the order in which the words are arranged.

The level of the word whose DEP is the root is set to 1, and if the DEP of the corresponding word corresponds to acl or relcl, 1 is added to the level of the head.

If the DEP of a word is not acl or relcl, the level of the word is set to the same level as that of the head. If the level of the head has not been determined yet, the level of the corresponding word is passed without setting, and if the last word is reached, the level is calculated again from the first word.

< Group decision >

The control module 110 groups words having the same set value into one box in FIG. 6 and sets the corresponding set value as an index box (index_box) value.

According to an embodiment of the present invention, the head index value of the words in each box is tracked, and a head index value of a word having a word other than a word in the box as a head is set as a head index box (head_index_box) value. For example, in FIG. 6, the head index value of 'a' and 'retired' among the words having a set value of 2 is 2, so the word 'soldier' in the same set (box) is the head, but the head index of 'soldier' is 11, so we have as head the word 'died' from another set (box). Accordingly, the head index box of index box 2 becomes 11.

When the DEP value of a word is acl, relcl, or root, the control module 110 sets the corresponding word as a group factor and sets a box including the corresponding word as a group factor_box.

In the case of the box corresponding to the group factor box, the index box value is set as the index group value, and when the head index box corresponds to the group factor box, the head index box value is set as the index group (index_group) value. For example, since the head index box 'died' of 'a retired soldier' with the index box value of 2 corresponds to the group factor box (true), the head index box value 11 is set as the index group value.

In the present invention, the frequency of sentences in foreign language textbooks and reference books is checked.

For example, content can be provided to learners by arranging words with the highest frequency among expressions in a geography application.

In addition, the system of the present invention checks whether the sentence has been learned, and the learned sentence is not presented to the learner.

Among the unlearned sentences, the sentence with the highest frequency is recommended.

Language is largely composed of subject part, relation part, object part (object), and background.

The subject part consists of a decorating part + subject, and the relational part consists of a decorating part + object.

In the present invention, the order of the subject part, the relation part, the target part (object), and the background is identified in the language to be learned, and the words of the Korean sentence are changed to the arrangement of the foreign language.

The present invention provides a language meta-format.

The present invention presents components that become the skeleton of a sentence after dividing all sentences into clauses and relatives.

Distinguish between key phrases and relatives based on the DEP values described above.

relative consists of 'acl', 'relcl', 'advcl', 'amod_t', 'appos', 'ccomp', 'csubj', 'csubjpass', and 'xcomp' components.

In addition, a bundle is formed according to the level value of a word or set.

When the level value is 1, it is treated as a key phrase, and when the level value is 2 or more, it is treated as relatives. If the level value is 3 or higher, all of them can be treated as level 2.

According to an embodiment of the present invention, the system first presents a sentence for the learner to select. The corresponding sentence may provide pronunciation through TTS.

The invention described above is not limited to the configuration and method of the embodiments described above, but all or some of the embodiments may be selectively combined so that various modifications may be made.

Claims (1)

The present invention provides a foreign language sentence structure learning system capable of learning foreign language sentence structures through a chatting method using a messenger.

Images