KR102171373B1 - Apparatus for korean morphemic analysis and method for the same - Google Patents

Abstract

It relates to a Hangul morpheme analysis apparatus and method, wherein the Hangul morpheme analysis apparatus determines whether or not a word to be analyzed is a homophonic participant based on a morpheme analysis database and the morpheme analysis database, and when the analysis target word is a homophone And a processor that determines at least one part-of-speech corresponding to the word to be analyzed based on the multiple part-of-speech determination function.

Description

Hangul morpheme analysis apparatus and method {APPARATUS FOR KOREAN MORPHEMIC ANALYSIS AND METHOD FOR THE SAME}

It relates to a Hangul morpheme analysis apparatus and method.

Text analysis is used in various fields. Text analysis analyzes the contents of various texts (e.g., posts, articles, replies, reviews and/or product evaluations such as literary works, social network services, etc.), and extracts valuable information from the text itself based on this It means performing statistical processing to obtain statistical analysis results. For example, companies or public institutions use text analysis to check consumer or public perceptions or public opinion through sentences written on the Internet. As a result, sites specializing in text analysis are emerging, and attempts to include the results of text analysis in the sites of each company or public institution are increasing.

The key to text analysis like this depends on the performance of morpheme analysis. This is because the meaning of the text may be different depending on the result of morpheme analysis. In other words, if the morpheme analysis is not properly performed, since it may be interpreted as a sentence with different intentions from the original intention, accuracy and consistency of the analysis may deteriorate.

Korean is one of the languages that morpheme analysis is very difficult. This is due to the nature of the agglutinating language. Agglutination is a language in which formal morphemes are added to real morphemes. Since two or more morphemes are fused into one syllable, the difficulty of separating morphemes is high, and the meaning domains of real morphemes and formal morphemes are different? However, there is a problem that different homophones frequently occur. In particular, among homophones, homophones having the same external form but different parts of speech are a major cause of deteriorating the performance of the morpheme analysis program by further increasing the difficulty of morpheme analysis. Therefore, effective extraction and processing of homophones in sentences such as this is inevitably required to increase the accuracy of morpheme analysis and overall sentence analysis.

By efficiently and appropriately determining the part of speech of a word or morpheme corresponding to a homoemic part of speech, it is intended to solve the problem of providing a Hangul morpheme analysis device and method that can improve the analysis performance of the text by analyzing the text more accurately. Make it an assignment.

In order to solve the above-described problems, a Hangul morpheme analysis apparatus, system, and method are provided.

Based on the morpheme analysis database and the morpheme analysis database, the Hangul morpheme analysis apparatus determines whether or not a word to be analyzed is a homophonetic part of speech, and when the word to be analyzed is a homophonic part-of-speech, among a single judgment function and a multiple part of speech judgment function It may include a processor that determines at least one part-of-speech corresponding to the word to be analyzed based on any one.

When it is determined that the analysis target word is not a homophonic part of the word, the processor may perform a general morpheme analysis of the analysis target word.

The morpheme analysis database may include a multi-part-of-speech root database constructed with at least one root that can be used as a plurality of parts of speech, and a multi-part-of-speech ending database constructed with at least one ending that can be used as a plurality of parts of speech.

The processor, if any one morpheme extracted from the analysis target word exists in the plural part-of-speech root database, and the other morpheme exists in the multiple part-of-speech mother database, the analysis target is based on a single part-of-speech determination function. Part of speech of at least one morpheme of a word can be determined.

The single part of speech determination function may include at least one of a general noun determination function, a proper noun determination function, a dependent noun determination function, a rhetorical determination function, a general verb determination function, an adjective determination function, an auxiliary verb determination function, and a root determination function. have.

The morpheme analysis database may further include a homophone type database including at least one type of a homogeneous product.

When the at least one morpheme extracted from the analysis target word does not exist in the multi-part-of-speech root database and the multi-part-of-speech ending database, the processor may browse the homoeum non-part of speech type database.

The processor may determine the part-of-speech of the at least one morpheme based on the multi-part-of-speech determination function based on a result of the at least one morpheme reading the homophoneic part-of-speech type database.

The multiple part-of-speech determination function may be obtained by combining at least two or more different single part-of-speech determination functions.

The different parts of speech determination functions may include at least two of a verb determination function, a noun determination function, and a root determination function.

The Hangul morpheme analysis method includes a step of inputting a sentence, separating a word from the sentence, determining whether or not the word to be analyzed is a homophone, based on a morpheme analysis database, and if the word to be analyzed is a homophone, It may include the step of determining at least one part-of-speech corresponding to the word to be analyzed based on any one of a single determination function and a plurality of parts of speech determination functions.

The Hangul morpheme analysis method may further include performing a general morpheme analysis on the analyzed word when it is determined that the analysis target word is not a homophonetic binary part of speech.

The morpheme analysis database may include a multi-part-of-speech root database constructed with at least one root that can be used as a plurality of parts of speech, and a multi-part-of-speech ending database constructed with at least one ending that can be used as a plurality of parts of speech.

If the analysis target word is a homophonetic part of speech, determining at least one part of speech corresponding to the analysis target word based on any one of a single determination function and a plurality of parts of speech determination function includes: If one morpheme exists in the plural part-of-speech root database and the other morpheme exists in the plural part-of-speech mother database, determining the part-of-speech of at least one morpheme of the word to be analyzed based on a single part of speech determination function Can include.

The single part of speech determination function may include at least one of a general noun determination function, a proper noun determination function, a dependent noun determination function, a rhetorical determination function, a general verb determination function, an adjective determination function, an auxiliary verb determination function, and a root determination function. have.

The morpheme analysis database may further include a homophone type database including at least one type of a homophone type database.

If the analysis target word is a homophonetic part of speech, determining at least one part of speech corresponding to the analysis target word based on any one of a single determination function and a plurality of parts of speech determination function includes at least one extracted from the analysis target word. If one morpheme does not exist in the plural part-of-speech root database and the plural part-of-speech ending database, the step of reading the homoemic part-of-speech type database may be included.

If the analysis target word is a homophonic part of speech, determining at least one part of speech corresponding to the analysis target word based on any one of a single determination function and a plurality of parts of speech determination function, wherein the at least one morpheme is the homophone The method may further include determining a part-of-speech of the at least one morpheme based on the multiple part-of-speech determination function based on a result of reading the part-of-speech type database.

The multiple part-of-speech determination function may be obtained by combining at least two or more different single part-of-speech determination functions.

The different parts of speech determination functions may include at least two of a verb determination function, a noun determination function, and a root determination function.

The Hangul morpheme analysis system includes a terminal device into which a sentence including an analysis target word is input and the sentence is received, the analysis target word is obtained from the sentence, and the analysis target word is homophoneized based on the morpheme analysis database. A Hangul morpheme analysis device that determines whether or not, and determines at least one part-of-speech corresponding to the word to be analyzed based on any one of a single decision function and a multiple part-of-speech decision function when the analysis target word is a homophonetic part of speech. Can include.

According to the above-described Hangul morpheme analysis apparatus and method, it is possible to simply, efficiently and appropriately determine the part-of-speech of a word or morpheme corresponding to a homophonetic part of speech, and accordingly, the entire text in the form of letters or voices can be more accurately analyzed. As a result, it is possible to improve the analysis performance for text.

According to the above-described Hangul morpheme analysis apparatus and method, it is possible to obtain an effect of more accurately and easily performing the morpheme analysis of the text of an awkward language (for example, Korean) that is very difficult to analyze the morpheme.

The above-described Hangul morpheme analysis apparatus and method can be used as a tool for the Hangul corpus in each field when analyzing articles, posts, comments, reviews and/or product evaluation, and/or public institutions or companies There is also an advantage that can be applied to various text analysis systems or programs, such as being able to be used as a core algorithm for text analysis in the case of conducting research or evaluation.

1 is a diagram of an embodiment of a Hangul morpheme analysis system.
2 is a block diagram of an apparatus for analyzing a Hangul morpheme according to an embodiment.
3 is a block diagram of an embodiment of a Dongeum Ipposa processing unit.
4 is a diagram showing a program code for an example of a single part of speech determination function.
5 is a diagram for explaining Hangul morpheme analysis.
6 is a diagram showing a program code for an example of a plurality of parts of speech determination function.
7 is a diagram for explaining an example of connection of a single part of speech determination function.
8 is a diagram illustrating a program code corresponding to an example of connection of a single part of speech determination function.
9 is a diagram for explaining an example of a multiple part of speech determination function.
10 is a diagram showing a program code corresponding to an example of a multiple part of speech determination function.
11 is a diagram showing a program code for an example of a verb-general noun function.
12 is a flowchart of an embodiment of a method for analyzing Korean morphemes.

In the following specification, the same reference numerals refer to the same elements unless otherwise specified. The term "unit" used below may be implemented as software or hardware, and according to an embodiment, the term "unit" is implemented as one part, or one "unit" is implemented as a plurality of parts. It is also possible.

When a part is said to be connected to another part throughout the specification, it may mean a physical connection depending on the part and another part, or may mean electrically connected. In addition, when a part includes another part, this does not exclude another part other than the other part unless otherwise stated, and it means that another part may be included further according to the designer's choice. do.

Terms such as first and second are used to distinguish one part from another, and unless otherwise specified, they do not mean sequential expressions. In addition, expressions in the singular may include plural expressions unless there is a clear exception in context.

Hereinafter, an embodiment of a Hangul morpheme analysis apparatus and a Hangul morpheme analysis system including the same will be described with reference to FIGS. 1 to 12.

1 is a diagram of an embodiment of a Hangul morpheme analysis system.

Referring to FIG. 1, the Hangul morpheme analysis system 1 may include a Hangul morpheme analysis apparatus 100 according to an exemplary embodiment. The Hangul morpheme analysis apparatus 100 is provided to perform a Hangul morpheme analysis from a sentence input from a user or another computing device (eg, the terminal device 20 or another server device).

The Hangul morpheme analysis apparatus 100 may be implemented based on one computing device (eg, a server device), or may be implemented using two or more computing devices. In the case of using two or more computing devices, the two or more computing devices may be connected to each other to be communicatively connected directly or through an external network 9. In addition, when using two or more computing devices, according to an embodiment, at least one computing device(s) stores a morpheme analysis database (111 in FIG. 2), and the other at least one computing device(s) is a processor ( It is also possible to be designed to perform the operation 130 of FIG.

The Korean morpheme analysis apparatus 100 connects to the network 9 and connects to at least one sentence in the form of text or voice (may be composed of a plurality of words or may be composed of only one word depending on the situation. Hada), or at least one sentence in the form of text or voice may be directly input from the user.

According to an embodiment, the Hangul morpheme analysis system 1 may further include at least one terminal device 10. The at least one terminal device 10 may receive, for example, a sentence in text or voice form and/or a result of analyzing a Hangul morpheme and output it to the outside.

The terminal device 10 may perform mutual communication with the Korean morpheme analysis apparatus 100 based on a wired communication network, a wireless communication network, or a combination thereof. Here, the wired communication network may be constructed using a cable, and the cable may be implemented using, for example, a pair cable, a coaxial cable, an optical fiber cable or an Ethernet cable. The wireless communication network may be implemented using at least one of a short-range communication network and a long-distance communication network. The short-range communication network is, for example, Wi-Fi, Wi-Fi Direct, zigbee, Bluetooth, Bluetooth Low Energy, CAN communication, and / Or it may be implemented using NFC (Near Field Communication). The telecommunication network may be implemented based on a mobile communication standard such as a wired communication network, for example, 3GPP, 3GPP2, or WiMAX series.

The terminal device 10 includes a cellular phone, a smart phone, a tablet PC, a desktop computer, a laptop computer, a smart watch, a head mounted display (HMD) device, a digital television, a set-top box, a navigation device, a portable game machine, a personal digital device. It may include a personal digital assistant (PDA), an electronic blackboard, an electronic billboard, and/or a sound reproduction device, but is not limited thereto.

Hereinafter, the Hangul morpheme analysis apparatus 100 will be described in more detail.

2 is a block diagram of an apparatus for analyzing a Hangul morpheme according to an embodiment.

Referring to FIG. 2, the Korean morpheme analysis apparatus 100 may include a storage unit 110 and a processor 130, according to an embodiment, and the input/output unit 101 or It may further include a communication unit 109.

The input/output unit 101 is provided to receive and/or output a sentence or various data from the outside. The input/output unit 101 may include at least one of the input unit 102 and the output unit 103. The input unit 102 is provided to receive a sentence, etc. from a user, for example, a keyboard, a keypad, a touch screen, a touch pad, a trackball, a track pad, a mouse device, a microphone, a stylus fan, a data input terminal And/or a motion detection sensor. The output unit 103 is provided to visually and/or audibly provide the analysis result of the Hangul morpheme to the user. The output unit 103 may include, for example, a display panel, a lighting device, a data output terminal, and/or a speaker device.

The communication unit 109 connects to the communication network 9 to access other devices (100A, 100B, etc.) in the Korean morpheme analysis device 100 and/or other external devices (for example, the terminal device 10, or an external device). It is provided to transmit data to a server device (not shown), and/or receive data from an external device. For example, the communication unit 109 may receive a sentence from the terminal device 10 in a predetermined form (eg, in text form).

The input/output unit 101 and/or the communication unit 109 are electrically connected to the processor 130 and transmit data to the processor 130 or receive data from the processor 130. If necessary, the input/output unit 101 and/or the communication unit 109 may be electrically connected to the storage unit 110 as well.

The storage unit 110 may store various types of data or applications required for the operation of the Korean morpheme analysis apparatus 100. For example, the storage unit 110 may store at least one application (which may be referred to as a program or an app) related to morpheme processing, homophony determination and processing, and/or result processing. Here, the application stored in the storage unit 110 may be previously created by the designer and stored in the storage unit 110, or obtained or updated through an electronic software distribution network accessible through a wired or wireless communication network 9 It may have been.

According to an embodiment, the storage unit 110 may store the morpheme analysis database 111. The morpheme analysis database 111 may be called by the processor 130 to be used for determination of a homophone or a process for a homophone.

The morpheme analysis database 111 may include, for example, at least one of a plurality of parts-of-speech root database 113, a plurality of parts-of-speech ending database 115, and a homoemic part-of-speech type database, as illustrated in FIG. 2.

The multiple part-of-speech root database 113 may be constructed with root(s) having multiple parts of speech. A root with plural parts of speech means a root that can be interpreted as two or more parts of speech in a sentence, and more specifically, any of a noun (country name) or a verb (think) within one sentence, such as [Koryo]. It means a root that can be used as one. Roots with such multiple parts of speech include, for example, Ga, Close, Strong, Strong, Gyeongmi, Grip, Coko, Tomorrow, Singing, Song, Glad, Daehan, Doyeon, Everytime, Momo, Banha, Librarian, Sasindo, In fact, there are Zinc, Ahead, Wait, Be, Ying, Ja, Zara, Acquaintances, Chae, Reference, Other, Ter, P.D., Hanjeok, Hal, Halgeo, Hwaha or Heegu. In addition to these, there are roots that have various parts of speech. It is estimated that about 1,690 roots have such multiple parts of speech.

The multiple part-of-speech root database 113 may be constructed including one or more roots having such a plurality of parts of speech. The multiple part-of-speech root database 113 may include roots having all the estimated multiple parts of speech, or may include only some roots of these.

According to an embodiment, the roots stored in the plural part-of-speech root database 113 may be provided to include verbs that do not change their form, such as body language, adverb, interjection, tubular sentence, or some roots.

The multiple part-of-speech ending database 115 may be constructed with ending(s) having multiple parts of speech. A ending having multiple parts of speech means a ending that can be judged as two or more parts of speech in a sentence. Specifically, it means a ending that can be interpreted as two or more parts of speech, such as investigation (-a), noun (distance), questionable ending (-ga?) or verb (to move), as in [A].

According to an embodiment, a ending stored in the multiple part-of-speech ending database 115 may include an affix, and may include a root in which affixed word form change occurs according to an embodiment.

Mothers having multiple parts of speech stored in the multiple part-of-speech mother database 115 are, for example, Ga, Branch, in the section, Guman, Me, I, because I am, I, I, and I. All, man-eun, man-in, envy, asa, child, oh, cry, shall we cry, the more we cry, the more we cry, the more, eun, yiga, jun, gave, give, give, give, give, give, give, give , Giving, giving, giving, giving, giving, giving, giving, giving, if giving, giving, giving, but giving, even if giving, Jundan, giving, Jundam, Junda, giving, Junda , Joon-Dani, Joon-dae-do, Joon-dae-yo, Joon-dae-yu, Joon-deul, Joon-d-ya, Joule, Joulgang, Jool-gul, I will give, I’ll give, I’ll give, I’ll give, I’ll give, I’ll give, I’ll give Can include, give, give, give, give, give, give, give, give, give, give, give, etc. In addition to these, there are mothers with various parts of speech. It is estimated that approximately 265 endings have such multiple parts of speech.

The above-described plural part-of-speech root database 113 and the plural part-of-speech ending database 115 may be called by the homophonic part-of-speech determination unit 133 of the processor 113, respectively, and may be used for the determination of a homophonic part of speech .

The homoeum i-Pomsa type database 117 may be constructed in the form of words or morphemes having homoeum i-Pomsa. For example, one morpheme may be used as a noun or a verb in a sentence, which is stored in the database 117 as a noun-verb type. In this case, a word or morpheme having a homoemic two parts of speech may be used as two types of parts of speech, or may be used as three or more parts of speech. The number of these types is estimated to be approximately 78.

Specifically, for example, the types stored in the homoemic participant type database 117 are nouns-interjections, nouns-adverbs, nouns-tubular criminals, nouns-dependent nouns, nouns-pronouns, nouns-rheutics, nouns-adjectives, nouns -Verbs, nouns-auxiliary verbs, nouns-roots, proper nouns-interjection, proper nouns-roots, proper nouns-subjectives, proper nouns-adverbs, proper nouns-adjectives, proper nouns-verbs, pronouns-verbs, pronouns-verbs- Adverbs, pronouns-root, rhetoric-root, adjective-adverb, adjective-root, verb-adverb, verb-adjective, verb-auxiliary verb, verb-root, root-adverb, interjection-root, tube detective-rhetoric, tube detective-root , Dependent noun-adverb, dependent noun-tubular criminal, dependent noun-verb, dependent noun-root, noun-adverb-root, noun-dependent noun-adverb, noun-dependent noun-adverb, noun-dependent noun-tubular criminal, noun- Dependent noun-pronoun, noun-dependent noun-adjective, noun-dependent noun-verb, noun-dependent noun-root, noun-pronoun-adverb, noun-pronoun-tubular detective, noun-pronoun-rheumatic, noun-pronoun-adjective, Noun-pronoun-verb, noun-pronoun-root, noun-adjective-root, noun-verb-root, noun-verb-adverb, noun-verb-adjective, noun-verb-auxiliary verb, noun-root-adverb, pronoun -Adverbs-root, pronoun-verb-adjective, pronoun-verb-auxiliary verb, pronoun-verb-root, pronoun-root-adverb, verb-adjective-adverb, verb-adjective-auxiliary verb, verb-auxiliary verb-adverb, Verb-root-adverb, noun-dependent noun-adverb-verb, noun-dependent noun-tubular criminal-verb, noun-dependent noun-pronoun-tubular criminal, noun-dependent noun-pronoun-verb, noun-dependent noun-adjective-auxiliary Pronouns, nouns-dependent nouns-verbs-adjectives, nouns-dependent nouns-verbs-auxiliary verbs, nouns-dependent nouns-verbs-roots, nouns-pronouns-verbs-roots, nouns-dependent nouns-pronouns-dictum-roots, nouns -Dependent noun-pronoun-verb-auxiliary verb, noun-dependent noun-pronoun-verb-root, noun-dependent noun-verb-adjective-auxiliary verb, noun-dependent noun-pronoun-tubular criminal-verb-auxiliary verb and/or Can include noun-dependent noun-pronoun-verb-adjective-root, etc. Depending on the embodiment, the homoeum I-POS type database 117 may include only some of these types, or may include other types in addition to these types.

The Dongeum IpoSa type database 117 may be called by the Dongeum Ipumsa processing unit 137 of the processor 130, and may be used for processing of the Dongeum Ipumsa.

The storage unit 110 described above may include at least one of a main memory device and an auxiliary memory device according to an embodiment. The main memory device may be implemented using a semiconductor storage medium such as ROM and/or RAM. The ROM may include, for example, a conventional ROM, EPROM, EEPROM, and/or MASK-ROM. The RAM may include, for example, DRAM and/or SRAM. Auxiliary storage devices include flash memory devices, SD (Secure Digital) cards, solid state drives (SSDs, Solid State Drives), hard disk drives (HDDs, Hard Disc Drives), magnetic drums, compact disks (CDs), DVDs ) Or an optical media such as a laser disk, a magnetic tape, a magneto-optical disk, and/or a floppy disk, and at least one storage medium capable of permanently or semi-permanently storing data.

The processor 130 may perform various operations, judgments, processing, and/or control operations required for the operation of the Hangul morpheme analysis apparatus 100. The processor 130 may drive an application stored in the storage unit 110 to perform such an operation.

Such a processor 130 is, for example, a central processing unit (CPU, Central Processing Unit), a microcontroller unit (MCU, Micro Controller Unit), a Micom (Micro Processor), an application processor (AP, Application Processor). , An electronic control unit (ECU) and/or other electronic devices capable of processing various operations and generating control signals. These devices can be implemented using, for example, one or more semiconductor chips and related components.

Referring to FIG. 2, the processor 130 includes a word separation unit 131, a homophonic part-of-satellite determination unit 133, a general morpheme processing part 135, a disequilibrium part-of-satellite processor 137, and a result acquisition and processing unit. (139) may be included. Some of the word separating unit 131, the homophoneous part of the word determination unit 133, the general morpheme processing unit 135, the disproportionate part of the word processing unit 137, and the result acquisition and processing unit 139 may be omitted as necessary. In addition, these 131 to 139 may be physically or logically classified according to an embodiment.

The word separation unit 131 may separate a word from a sentence. A word is the smallest unit of a sentence component and consists of one or more words or morphemes. The word may be input from the input unit 102 and/or received from the communication unit 109. In this case, the storage unit 110 first stores the input and/or received sentences, and the word separation unit 131 calls and calls the sentence stored in the storage unit 110 in response to an input such as a homoemic part of speech processing command. Word separation can also be performed on a sentence. One or two or more words separated from the sentence may be transmitted to the homophonetic part-of-speech determination unit 133.

The homoeum i-Pomsa determination unit 133 determines whether a word corresponds to a homoeum i-Pomsa, or extracts a morpheme (which may include a real morpheme or a formal morpheme) from the word, and then determines whether the extracted morpheme is a homoemic part of the body. I can judge.

According to an embodiment, the homoemic part of the word determination unit 133 may determine whether a word corresponds to a homogeneous part of the word based on the morpheme analysis database 111. For example, the homophonetic part-of-speech determination unit 133 uses the plural part-of-speech root database 113 and/or the plural part-of-speech ending database 115 among databases 113, 115, and 117 of the morpheme analysis database 111 It is possible to determine whether this homogeneity corresponds to the POS.

Specifically, for example, the homophonetic heterogeneous part of speech determination unit 133 first refers to the plural part-of-speech root database 113, and data corresponding to the received word or morpheme corresponding to the root of the word (word, morpheme, etc.) It is determined whether or not is present in the plural part-of-speech root database 113, and by sequentially referring to the plural part-of-speech ending database 115, data corresponding to the received word or morpheme corresponding to the ending in the word is a plurality of parts of speech ending database ( 115), it is also possible to determine whether or not the word corresponds to a homogeneous part of the word. In this case, when at least one of a word, a root, and a ending exists in the multiple part-of-speech root database 113 or the multiple part-of-speech root database 113, the homoemic heterogeneous part-of-speech determination unit 133 says that the word corresponds to the same-eum-different part of speech. I can judge.

If it is determined that a word or a morpheme included therein does not correspond to a homoemic part of the word, the word or morpheme is transmitted to the general morpheme processing unit 135, and the word or morpheme is analyzed according to a conventional morpheme analysis method. That is, on the premise that it is not a homogeneous part-of-sight, an unavoidable or morpheme analysis is performed according to at least one of conventionally known analysis methods.

On the contrary, if it is determined that a word or a morpheme included therein is a homoeum-ipomsa, the word or a morpheme included therein is transferred to the homoeum-ip-of-sa processing unit 137, and the homoeum-iso-Pomsa processing unit 137 performs an analysis on the word or morpheme. can do.

3 is a block diagram of an embodiment of a Dongeum Ipposa processing unit.

As shown in FIG. 3, after checking the multiple part-of-speech root database 113 (137a), and checking the multiple part-of-speech mother database 115 (137b), in one embodiment, , Based on the confirmation result, a morpheme analysis may be performed using a single part-of-speech determination function (137c) or a homophoneic part-of-speech type database 117 may be checked (137d). In the case of checking the homoeum two parts of speech type database 117 (137d), the homoeum two parts of speech processing unit 137 performs a morpheme analysis using a plurality of parts of speech determination function based on the confirmation result of the homoeum two parts of speech type database 117 Can perform (137e)

More specifically, the homoeum-iponics processing unit 137 may extract a root from the word and determine whether the extracted root exists in the multiple part-of-speech root database 113 (137a). In this case, the homoeum-ipomsa processing unit 137 may extract the root by sequentially reading from the first part of the word (the left part according to the Hangul description method). If the root of the word does not exist in the multiple part-of-speech root database 113, the homoeum-ip-of-satellite processing unit 137 checks the homoeum-ip-of-sa-type database 117.

On the contrary, if the root exists in the multiple part-of-speech root database 113, the homoeum 2 part-of-speech processing unit 137 may check the multiple part-of-speech ending database 115 (137b). In detail, the homoeum two parts of speech processing unit 137 may extract a ending from a word and determine whether the extracted ending exists in the multiple part of speech ending database 115 (137b). In this case, the homoeum two parts processing unit 137 extracts the ending while reading from the last part of the word (the right end part according to the Hangul writing method) to the first part of the word (the leftward direction according to the Hangul writing method). May be. If the mother does not exist in the multiple part-of-speech mother database 115, the homoeum I-POS processing unit 137 checks the dong-eum I-POS type database 117.

On the contrary, if the ending is present in the multiple part-of-speech ending database 115, the homophonic part-of-speech processing unit 137 performs morpheme analysis using a single part-of-speech determination function (137c).

3 is only shown for an embodiment in which the multi-part-of-speech root database 115 is checked after the multi-part-of-speech root database 113 is checked (137a), but the verification process (137a) of the multi-part-of-speech root database 113 and the multi-parts of speech The confirmation process 137b of the mother database 115 may be performed in an order opposite to that shown in FIG. 3. In addition, according to an embodiment, the verification process 137a of the multiple POS root database 113 and the verification process 137b of the multiple POS mother database 115 may be performed simultaneously.

The single part-of-speech determination function refers to a function capable of determining a specific part of speech (for example, a noun, pronoun, rhetoric, verb, adjective, coronal sentence, adverb, investigation, or interjection). When a word, word, or morpheme is input, the single part-of-speech determination function may determine whether the word, word, or morpheme corresponds to a specific part of speech (for example, a noun, etc.), and output a determination result. In this case, the single part-of-speech determination function may output True or False to determine whether a word, word, or morpheme corresponds to a specific part of speech. The single part-of-speech determination function can be variously determined according to the part of speech to be determined. For example, the single part-of-speech determination function includes a general noun determination function, a proper noun determination function, a dependent noun determination function, a rhetorical determination function, a general verb determination function, an adjective determination function, an auxiliary verb determination function, and/or a root determination function. Can include.

4 shows a program code for an example of a single part of speech determination function, and illustrates an example of a general noun determination function.

As illustrated in FIG. 4, the single part-of-speech determination function p1 may determine whether a specific part-of-speech exists in front of one word of the current morpheme to be examined (p11). In this case, if an adjective exists in front of the current word, it can be designed to determine the current morpheme as a noun. Such a judgment is not necessarily correct because adjectives can come after adjectives, such as in a clean, pretty and fine house. However, since this judgment is used when it is suspected that a specific form is a noun, misanalysis generally does not occur in situations where this rule is used.

In addition, the single part-of-speech determination function p1 may determine whether the current word ends with a vocabulary included in the first character string w11 (words) (p12). In general, nouns take investigations, so it is not unreasonable to make judgments based on them. However, since [silver] can be used as an auxiliary (for example, a patent, etc.) and as a ending (for example, thin, etc.), it may cause an error in morpheme analysis, so the first character string w11 It may be designed to include only cases that are used for investigation only.

In addition, the single part-of-speech determination function p1 may also determine whether the current word ends with a vocabulary included in the second character string w12 and words2. Certain morphemes (eg, [as much] or [like]) can only take nouns in front of them, so a general noun can be determined based on this.

On the other hand, as shown in Figure 3, if the root does not exist in the plural part-of-speech root database 113 or if the ending does not exist in the plural part-of-speech ending database 115, the homophones processing unit 137 is The type database 117 can be checked (137d). In this case, the homoeum two parts of speech processing unit 137 extracts at least one homoemic two parts of speech type from the homoeum two parts of speech type database 117, and analyzes the morpheme using a plurality of parts of speech determination function based on the extracted homoemic two parts of speech type Can be performed (137e).

5 is a diagram for explaining Hangul morpheme analysis.

For example, referring to the bar shown in FIG. 5, the morpheme of [fall] may have at least four meanings according to the context of a sentence. Specifically, in the sentence [I looked at a wide descent from a hill] (hereinafter, the first sentence), [descent] is a noun meaning river and river, and the sentence [I experienced a descent from the sky on a parachute] (hereinafter referred to as the first sentence). In sentence 2), [descend] is a noun that means to descend (fall). In addition, in the sentence [He is very strong] (hereinafter, the third sentence), [descend] is used as an adjective that means force is counting, and in the sentence [he descended below the cliff] (hereinafter, the fourth sentence) goes down ( Descend) is used as the root of meaning. As described above, the same morpheme may be determined to have similar or different meanings, depending on the context within a sentence, and may also be determined as different parts of speech.

In this case, when [descend] is used as a noun as in the first sentence and the second sentence, it often appears as an investigation, and if it is used as an adjective as in the third sentence, it is usually accompanied by the ending [-c]. It is often used. In addition, when used as a root as in the fourth sentence, there are many cases where [-hada]/[-become]/[-georida] is added after the root. Considering this point, it is possible to appropriately determine the part of speech of [Descent] and add additional information (ie, information about the part of speech of [Descent]) to the text. In this way, adding a part-of-speech corresponding to a certain morpheme to a certain morpheme is called part-of-speech tagging.

6 is a diagram showing a program code for an example of a plurality of parts of speech determination function.

As described above, when one word or morpheme can be used as a plurality of parts of speech, in order to implement part-of-speech tagging for one word or morpheme, a multiple part of speech determination function is required. The multiple part-of-speech determination function refers to a function capable of determining one of a word or morpheme that can be used as a plurality of parts of speech.

As shown in FIG. 6, the multiple part of speech determination function may be written and implemented in program code p2. For example, a multi-part-of-speech determination function defined as [fall] is implemented to determine the part of speech of the word [fall] within a word or morpheme. In more detail, for example, the multi-part-of-speech judgment function defined as [Descent] is provided so that the word to be analyzed (hereinafter, the word to be analyzed) is input as a variable, and the word to be analyzed is [i] , [Is] or [in], etc. (p21), or [i], [a] or [b], etc. (p22), the part-of-speech of [descending] within a word or morpheme is used as a noun. If the words to be analyzed include [C], [Shou], or [Come] (p23), the part of speech of [Descent] should be judged as an adjective. ] Or [Hani, etc.] (p24) may be designed to determine the part of speech of [Descent] as a root.

7 is a diagram for explaining an example of connection of a single part of speech determination function, and FIG. 8 is a diagram illustrating a program code corresponding to an example of connection of a single part of speech determination function.

However, building the program code for all possible words (for example, words included in the multi-part-of-speech root database 113 in addition to [Descent]) increases the amount of data and reduces performance accordingly. Maintaining code consistency also becomes difficult. Therefore, as shown in Figs. 7 and 8, if a single part of speech function is constructed and then each of the constructed single parts of speech functions are connected to determine a specific part of speech, the amount of data is reduced, the code is simplified, and the code is updated. It is possible to achieve consistency of

Specifically, as shown in FIG. 7, a single part-of-speech determination function p3a, p3b, and p3c for each of a plurality of parts of speech may be interconnected and used. Here, the single part of speech judgment function (p3a, p3b, p3c) is a general noun judgment function, a proper noun judgment function, a dependent noun judgment function, a rhetoric judgment function, a general verb judgment function, an adjective judgment function, an auxiliary verb judgment function, and a root judgment function. It may include at least two of.

Specifically, for example, the verb determination function p3a and the noun determination function p3b are connected and used to determine the part-of-speech of the word A (W11) belonging to the homophonetic part of speech. In other words, the verb judgment function (p3a) and the noun judgment function (p3b) can be connected and implemented as a single function, and the word A (w11) is substituted into the function to which these (p3a, p3b) are connected to be used as a verb or a noun. Is determined, and a determination result R11 is obtained accordingly. Similarly, the noun determination function (p3b) and the root determination function (p3c) can be connected to each other and implemented as a single function, which can be used to determine the word B (w11) belonging to different homophones as a noun or root. Can (R12).

When at least two of the plurality of single parts of speech determination functions p3a, p3b, and p3c are interconnected and used, the program code p3 may be written as shown in FIG. 8. In this case, when compared with the program code p2 shown in FIG. 6, the program code p3 is relatively simplified by modularizing the parts P21 and P22 related to the noun determination (P31).

9 is a diagram for explaining an example of a multiple part of speech determination function.

As shown in FIGS. 7 and 9, when single parts of speech functions for each word are connected and used, the amount of code is drastically reduced, and consistency and ease of management of the code can be expected. However, since the number of homophones to be functionalized is large, the code to be written is still not short. For example, words used as nouns, adjectives, and roots all have a problem that the same code used in [Descent] must be used repeatedly.

Therefore, in order to reduce the amount of code, automate tasks, and maintain the consistency of the code accordingly, as shown in FIG. 9, by interconnecting two or more single part-of-speech determination functions f1 to f3, at least one multiple part of speech Decision functions (f12, f23, f123, etc.) can be implemented.

Specifically, a judgment function (f1, hereinafter part of speech A judgment function) for part of speech A (for example, a verb), a judgment function (f2, hereinafter part of speech B judgment function) for part of speech B (for example, noun), and When a determination function (f3, hereinafter, a part of speech C determination function) for a part of speech C (for example, a root) is given, a function capable of determining a plurality of parts of speech may be designed and constructed based on this. For example, a part-of-speech A decision function (f1) and a part-of-speech B decision function (f2) are combined and connected to obtain a part-of-speech AB decision function (f12, for example, a verb-noun decision function), or a part of speech B decision function ( By combining and connecting f2) and the part-of-speech C determination function f3, a part-of-speech BC determination function f23 (eg, a noun-root determination function) may be obtained. In addition, the part-of-speech ABC determination function (f123, for example, verb-noun-root determination function) by combining and combining all of the parts of speech A determination function (f1), the part of speech B determination function (f2), and the part of speech C determination function (f3) ) Can also be obtained. At least one of these judgment functions (f12, f23, f123, etc.) is at least one word that can be used as a plurality of parts of speech (e.g., word A(w1), word B(w2) and the word C(w3), etc.) may be applied respectively, and accordingly, the part-of-speech of at least one word (w1, w2, w3) may be determined.

10 is a diagram showing a program code corresponding to an example of a multiple part of speech determination function.

When at least one multi-part-of-speech determination function f12, f23, and f123 is implemented by combining a plurality of single part-of-speech determination functions f1 to f3 as described above, determining the part-of-speech of [falling] shown in FIG. 6 or 7 As shown in FIG. 10, the function for may be written in a simpler and simpler program p4. In this case, a single line of simple function call statements (p41, for example, a call statement of the verb-noun-root determination function f123) can analyze parts of speech according to the context. For example, descent can be analyzed as one of a noun, an adjective, and a root. Accordingly, conciseness and consistency of code can be implemented more effectively.

11 is a diagram showing a program code for an example of a verb-general noun function.

The program code p4 shown in FIG. 10 may call a verb-general noun function P5 as shown in FIG. 11, for example. In this case, the called verb-general noun function P5 may be implemented by combining several decision functions p51 to p55, for example. More specifically, the verb-general noun function (P5) is a dependent noun determination function (p51), a general verb determination function (p52), an adjective determination function (p53), a general noun determination function (p54), and/or other determination functions (p55). ) May be included, and may be implemented by sequentially combining these (p51 to p55).

In more detail, the dependent noun function p51 may be used to determine whether the part of speech of a word to be analyzed or a morpheme included therein is a noun. The dependent noun function p51 may check whether a rhetoric (number) exists in a word (or morpheme, etc.) to be analyzed currently or a word (or morpheme, etc.) of a front end of the word to be analyzed. The number may be expressed in Korean, Chinese characters, alphabets, or Arabic. In the case of having to decide on either a noun or a verb, if there is a rhetoric in the word preceding the word to be analyzed, it is very likely that it is the subject to be analyzed. The dependent noun function p51 preferentially processes the case where it is almost certain in this way.

In addition, a verb determination function p52 and an adjective determination function p53 may be used to determine whether the part of speech of a word to be analyzed or a morpheme included therein is a verb. The determination of whether or not to be a verb may be coded to be processed prior to the determination of whether or not to be a noun using the noun determination function p54. This is because the characteristics of verbs are relatively clearer in situations in which verbs or nouns must be judged. Specifically, in the case of the sentence [Take off the paper], if you first determine whether or not it is a noun, there is a high likelihood that [they] will be judged as a noun. This is because nouns are well combined with the terminating ending [-c] (for example, he is a doctor[da]). However, since verbs have the characteristic of having an object or adverb before it, if you first check whether an object or adverb is present in the word before the word to be analyzed, then [Flock] in [Tear Off Paper] is appropriately used as a verb. Can be determined. Verb determination may be performed based on a verb determination function p52 which is a single part of speech determination function. Adjectives can also be preceded by an adverb, so they can be judged in advance of a noun. Determination of adjectives may also be performed using the adjective determination function p53.

Whether or not it is a noun may be checked sequentially using the noun determination function p54. The noun determination function p54, as illustrated in FIG. 6, has a form in which a specific part of speech or a specific form comes together, and thus can be determined based on this.

In addition, the verb-general noun function P5 may include a function p55 for determining whether the length of the word or morpheme in the word to be analyzed is the same as the length of the word to be analyzed. If a word or morpheme exists alone in a word (in other words, if the word has only one word or only one morpheme), then such a word or morpheme is usually a noun, very rarely or not a verb. . Therefore, in cases where both nouns and verbs can be judged, the verb-general noun function (P5) compares the length of a word or morpheme with the length of a word, and if the length is the same, it is judged as a noun, otherwise Try to judge by verb. Accordingly, a word or morpheme that corresponds to a homoeum-ipomsa can be determined as either a verb or a noun.

The result acquisition and processing unit 139 receives a processing result from at least one of the general morpheme processing unit 135 and the homoemic part of the speech processing unit 137, and obtains a final analysis result of the word of the input sentence. For example, if the general morpheme processing unit 135 and the homogeneous part-of-speech processing unit 137 acquire only the result of the determination of part of speech, the result acquisition and processing unit 139 is based on the obtained part of speech determination result, for each word. You can also do tagging. In addition, the result acquisition and processing unit 139 transmits the analysis result to the terminal device 10 or the like, or transmits the analysis result to the storage unit 110 to be stored, and/or the analysis result is transmitted to the output unit 103. It can also be controlled to be output to the outside through. In addition, the result acquisition and processing unit 139 may perform various arithmetic processing and/or control operations based on the analysis result according to the designer's arbitrary selection.

Hereinafter, an embodiment of a method for analyzing Korean morphemes will be described with reference to FIG. 12.

12 is a flowchart of an embodiment of a method for analyzing Korean morphemes.

According to an embodiment of the Hangul morpheme analysis method shown in FIG. 12, first, a sentence is input to a Hangul morpheme analysis apparatus or a terminal device communicatively connected thereto (151). The input of such a sentence may be performed directly by the user and/or may be performed by another computer device. In addition, the input of the sentence may be input in the form of text or voice. When a sentence is input in the form of voice or the like, the sentence in the voice form may undergo a conversion process to text.

The Hangul morpheme analysis apparatus may separate a word from a sentence directly input or transmitted from a terminal device (152). Depending on the embodiment, the terminal device may separate words from sentences. In this case, the Hangul morpheme analysis device receives the word.

The Hangul morpheme analysis apparatus may determine whether a word in a word, a word in a word, or a morpheme in a word corresponds to a homophonetic part of the word (153). The determination of the homoeum two parts of speech may be performed using, for example, a morpheme analysis database, and more specifically, may be performed using a multiple part-of-speech root database and/or a multiple part-of-speech ending database.

If it is determined that the homophone does not correspond to the part of the word (No in 153), the Hangul morpheme analysis device performs a commonly known general morpheme analysis on a word, a word within a word, or a morpheme within a word (154).

On the contrary, if it is determined that it corresponds to a homophoneous part of the story (example of 153), the Hangul morpheme analysis device may proceed with the analysis of the same sound part of the part (155). In this case, the Hangul morpheme analysis apparatus may perform an analysis on the homophones by using the morpheme analysis database. Here, the morpheme analysis database may include, for example, a multi-part-of-speech root database, a multi-part-of-speech ending database, and/or a homophonetic two-parts type database.

According to an embodiment, in order to perform a homophonetic part-of-speech analysis, first, it may be determined whether data corresponding to a root of a word exists in a multi-part-of-speech root database. If data corresponding to the root of a word or the like is retrieved from the multi-part-of-speech root database, it may be determined whether data corresponding to the ending of the word or the like exists in the multi-part-of-speech ending database. If data corresponding to a ending or the like can be obtained from a multiple part-of-speech ending database, a morpheme analysis is performed on a word or morpheme corresponding to a homophonic part-of-speech using a single part of speech determination function. According to an exemplary embodiment, the determination as to whether data corresponding to a ending or the like is detected from the multi-part-of-speech mother database may be performed prior to determining whether data corresponding to a root or the like is detected from the multi-part-of-speech root database. Also, both may be performed simultaneously.

In addition, if the data corresponding to the root of a word does not exist in the multi-part-of-speech root database, and/or data corresponding to the ending, etc. does not exist in the multi-part-of-speech ending database, the homophonic two parts of speech determination function is used. A morpheme analysis may be performed for a word or morpheme corresponding to. In this case, the homoeum heterophasic type database may be searched first, and at least one type of homoemic heterophony is identified according to the search result, and a morpheme analysis using a multi-part of speech determination function is performed based on the identified type of homoemic heterophony. It could be.

An analysis result according to performing a general morpheme analysis (154) or a homoemic part-of-satellite analysis (155) is sequentially obtained (156). The analysis result may be obtained after the general morpheme analysis 154 or the homoemic part-of-sight analysis 155 is finished. In addition, the analysis result may be obtained in real time in the process of performing a general morpheme analysis 154 or a process 155 of performing a homophonetic analysis. The obtained analysis result may be additionally processed or transformed according to the purpose of using the analysis result. In addition, it may be transmitted to a terminal device or provided to a user through a speaker or a monitor device connected to the Hangul morpheme analysis device.

The Hangul morpheme analysis method according to the above-described embodiment may be implemented in the form of a program that can be driven by a computer device. Here, the program may include a program command, a data file, a data structure, or the like alone or in combination. The program may be designed and produced using machine code or high-level language code. The program may be specially designed to implement the above-described method, or may be implemented using various functions or definitions previously known to and available to those skilled in the computer software field. In addition, here, the computer device may be implemented including a processor or a memory that enables the function of a program to be realized, and may further include a communication device if necessary.

A program for implementing the above-described method for analyzing Korean morphemes may be recorded in a recording medium readable by a computer. The computer-readable recording medium includes, for example, a magnetic disk storage medium such as a hard disk or a floppy disk, a magnetic tape, an optical recording medium such as a compact disk or a DVD, a magnetic-optical recording medium such as a floppy disk, and a ROM. , A semiconductor storage device such as RAM or flash memory, etc., may include various types of hardware devices capable of storing a specific program executed according to a call from a computer.

Although various embodiments of the Hangul morpheme analysis apparatus, system, and method have been described above, the Hangul morpheme analysis apparatus, system, and method are not limited to the above-described embodiments. Various devices, systems, or methods that can be implemented by modifying and modifying based on the above-described embodiment by a person of ordinary skill in the art may also be an embodiment of the aforementioned Hangul morpheme analysis device, system, or method. For example, the described techniques are performed in an order different from the described method, and/or components such as systems, structures, devices, circuits, etc. described are combined or combined in a form different from the described method, or other components or Even if it is substituted or substituted by an equivalent, this may be an embodiment of the apparatus, system, and method for analyzing the Hangul morpheme described above.

10: terminal device 100: Hangul morpheme analysis device
101: input/output unit 109: communication unit
110: storage unit 111: morpheme analysis database
113: Multiple Part of Speech Roots Database
115: multiple parts of speech mother database
117: Dongeum Part-Time Type Database
130: processor 131: word separation unit
133: Hom Eum Transplantation Judgment 135: General Morphological Processing Department
137: homoeum transfer part processing unit 139: result acquisition and processing unit

Claims (21)

Morpheme analysis database; And
Based on the morpheme analysis database, it is determined whether or not the word to be analyzed is a homophonic part-of-speech, and when the word to be analyzed is a homophonic part-of-speech, the analysis target word is based on any one of a single part of speech determination function and a multiple part of speech determination function Includes; a processor for determining at least one part of speech corresponding to,
The morpheme analysis database,
A multi-part-of-speech root database constructed with at least one root that can be used as a plurality of parts of speech; And
Includes; a plurality of parts of speech ending database constructed with at least one ending that can be used as a plurality of parts of speech,
The processor,
If any one morpheme extracted from the analysis target word exists in the plural part-of-speech root database, and the other morpheme exists in the multiple part-of-speech ending database, at least one of the analysis target word is based on a single part-of-speech decision function Hangul morpheme analysis device that judges the part of speech of morphemes of. The method of claim 1,
The processor, when it is determined that the analysis target word is not a homophonetic part-of-sight, performs a general morpheme analysis on the analysis target word. delete delete The method of claim 1,
The single part of speech determination function includes at least one of a general noun determination function, a proper noun determination function, a dependent noun determination function, a rhetorical determination function, a general verb determination function, an adjective determination function, an auxiliary verb determination function, and a root determination function. Morphological analysis device. The method of claim 1,
The morpheme analysis database,
Hangeul morpheme analysis apparatus further comprising a; homoeum i-Pomsa type database including at least one type of homoeum i-Pomsa. The method of claim 6,
The processor, when at least one morpheme extracted from the analysis target word does not exist in the multi-part-of-speech root database and the multi-part-of-speech ending database, the Hangul morpheme analysis apparatus for reading the homophoneic two-parts type database. The method of claim 7,
The processor, the Hangul morpheme analysis apparatus, wherein the at least one morpheme determines the part of speech of the at least one morpheme based on the plurality of parts-of-speech determination function based on a result of reading the homophoneic part-of-speech type database. The method of claim 8,
The plurality of parts-of-speech determination functions are obtained by combining at least two or more different single part-of-speech determination functions. The method of claim 9,
The different parts of speech determination functions include at least two of a verb determination function, a noun determination function, and a root determination function. As a Hangul morpheme analysis method performed by a Hangul morpheme analysis device,
Inputting a sentence;
Separating a word from a sentence;
Determining whether or not the analysis target word is a homoeum part of speech based on a morpheme analysis database; And
If the analysis target word is a homophonic part of speech, determining at least one part of speech corresponding to the analysis target word based on any one of a single part of speech determination function and a plurality of parts of speech determination function; and
The morpheme analysis database,
A multi-part-of-speech root database constructed with at least one root that can be used as a plurality of parts of speech; And
Includes; a plurality of parts of speech ending database constructed with at least one ending that can be used as a plurality of parts of speech,
If the analysis target word is a homophonic part of speech, determining at least one part of speech corresponding to the analysis target word based on any one of a single determination function and a plurality of parts of speech determination function,
If any one morpheme extracted from the analysis target word exists in the plural part-of-speech root database, and the other morpheme exists in the multiple part-of-speech ending database, at least one of the analysis target word is based on a single part-of-speech decision function Hangul morpheme analysis method comprising a; determining the part of speech of the morpheme of. The method of claim 11,
If it is determined that the analysis target word is not a homophonetic part of the word, performing a general morpheme analysis on the analysis target word; Hangul morpheme analysis method further comprising. delete delete The method of claim 11,
The single part of speech determination function includes at least one of a general noun determination function, a proper noun determination function, a dependent noun determination function, a rhetorical determination function, a general verb determination function, an adjective determination function, an auxiliary verb determination function, and a root determination function. Morphological analysis method. The method of claim 11,
The morpheme analysis database,
A method for analyzing Korean morphemes further comprising a homoeum iPomsa type database including at least one type of homoeum iPomsa. The method of claim 16,
If the analysis target word is a homophonic part of speech, determining at least one part of speech corresponding to the analysis target word based on any one of a single determination function and a plurality of parts of speech determination function,
When at least one morpheme extracted from the analysis target word does not exist in the multi-part-of-speech root database and the multi-part-of-speech ending database, reading the homophoneic two-parts type database; and Hangul morpheme analysis method comprising: The method of claim 17,
If the analysis target word is a homophonic part of speech, determining at least one part of speech corresponding to the analysis target word based on any one of a single determination function and a plurality of parts of speech determination function,
The method of analyzing Korean morphemes further comprising: determining, by the at least one morpheme, the part-of-speech of the at least one morpheme based on the plurality of parts-of-speech determination function based on a result of browsing the homophoneic part-of-speech type database. The method of claim 18,
The plurality of parts-of-speech determination functions are obtained by combining at least two or more different single part-of-speech determination functions. The method of claim 19,
The different parts of speech determination functions include at least two of a verb determination function, a noun determination function, and a root determination function. delete

Images