공형태소와 형태 분석에 대하여

시정곤 국어학회 2010 국어학 Vol.0 No.57

This paper explored the definition of a morpheme and its basic relational notions such as morpheme analysis, fossil, lexicalization etc. focusing on so called 'empty morphemes'. Regarding the empty morphemes, we examined the previous works again and emphasized that morphemes are the minimal units based on their meanings and empty morphemes are unacceptable in Korean grammar. This paper argued when we analyze morphemes, it must be analyzed only in synchronic grammar. This paper also argued that although morpheme analysis of words could be more detailed for maximizing explanation, we cannot regard the result as a morpheme in synchrony. We recognized an unit that cannot be analyzed in synchronic grammar as a fossil, although it was a morpheme in the past grammar. and the whole word that cannot be analyzed in synchronic grammar as a lexicalized word. In addition, in this paper, we argued that to have synchronic productivity was one thing, to be a synchronic morpheme was quite another thing. On the basis of this assumption, we suggested three types of morphemes such as formative, constructive, etymologic unit.

Morpheme Segmentation and Concatenation Approaches for Uyghur LVCSR

Mijit Ablimit,Tatsuya Kawahara,Askar Hamdulla 보안공학연구지원센터 2015 International Journal of Hybrid Information Techno Vol.8 No.8

In this paper, various kinds of sub-word lexica are thoroughly investigated under the framework of Uyghur LVCSR system. Experimental results show that it is inefficient to directly model based on word units or small units like morpheme or even syllable units. It is observed that an optimal sub-word unit set between word and morpheme units can better fit for ASR system. In order to select best unit set we have investigated several effective unit segmentation, concatenation approaches, and their ASR performances. For segmentation approach, we investigate a supervised segmentation which split words into the smallest functional units - the linguistic morphemes, and an unsupervised segmentation which extract pseudo-morphemes (or statistical morphemes). In supervised model, a leaning algorithm is trained on a manually prepared training corpus, and morpho-phonetics changes are analyzed. In the unsupervised model, the Morfessor tool is used to extract pseudo-morphemes from a raw text corpus. For concatenation approach, several effective concatenation approaches are investigated based on linguistic morphemes. First is the data-driven approach which concatenates morpheme sequences based on certain measures like co-occurrence frequency or mutual probability. Second is a model based approach which merges units with global statistical criteria. In this study, the Morfessor program is revised and turned into concatenation program by controlling segmentation points. Third is the two-layer-lexica based concatenation approach which extracts an optimal sub-word unit set by aligning and comparing the ASR results of word and morpheme two lexical layers. This method utilizes both speech and text, and produced the best results in terms of WER and lexicon size, and proved to be very stable. The best optimal lexicon, which is obtained totally on the basis of HMM based acoustic model, outperformed all other baseline lexica. And when all these lexica are directly incorporated with a deep neural network (DNN) based acoustic model, without changing the speech and text training corpora and language models, the optimal lexicon not only drastically improved the ASR accuracy but also outperformed other units as a proof of the generality of the two-layer-lexica based approach.

형태소의 식별과 분류

이선웅 ( Yi Seon-ung ),오규환 ( Oh Gyu-hwan ) 국어학회 2017 국어학 Vol.81 No.-

이 글에서는 형태소 식별(분석) 기준이 연구 목적에 따라 달라질 수 있다는 관점을 견지하면서, 형태소를 ‘자립형태소’와 ‘의존형태소’, ‘실질형태소’와 ‘문법형태소’로 분류할 때에 발생하는 문제들을 검토하였다. 이 글에서 논의된 내용을 정리하면 다음과 같다. 첫째, 형태소는 연구 목적과 연구 대상에 따라 달리 식별될 수 있으며 심리어휘부를 전제하는지 이론어휘부를 전제하는지에 따라서도 달리 식별될 수 있다. 둘째, 한국어의 형태소 분석은 IA 모형을 통하여 단어형으로부터 형태소를 식별하는 것으로부터 출발하며 단어형으로부터 형태소를 식별할 때에는 ‘결합관계’, ‘계열관계’, ‘음운론적 현현방식’ 등을 위계적으로 고려하여야 한다. 셋째, 공형태(소) 및 영형태소는 국부적으로는 유용한 개념일 수 있고 이론어휘부를 전제하는 연구에서도 필요한 개념일 수 있다. 다만, 심리어휘부와 이론어휘부의 일치를 추구하는 형태론에서라면 성립하기 어려운 개념이다. 넷째, 언어의 화석화와 공형태의 상관관계를 고려하면, 공형태와 화석은 ‘공형태이면서 화석인 것’, ‘공형태이지만 화석이 아닌 것’, ‘화석이지만 공형태가 아닌 것’으로 교차 분류할 수 있다. 다섯째, 의존성은 크게 ‘형태적 의존성’, ‘통사적 의존성’, ‘의미적 의존성’으로 나눌 수 있는데 이러한 의존성이 어떻게 나타나는지를 고려하여 전형적인 의존형태소부터 전형적인 자립형태소에 이르는 척도를 상정할 수 있다. 여섯째, 한국어 형태소의 실질성 위계는 잠정적으로 ‘[실질어휘] > [접두사 > 형식명사/보조용언/의존용언] > [보조사(양태적 의미 표현)/양태어미/일부 접미사 > 보조사(문법적 의미 표현)] > [의미격조사] > [문법격조사 > 접속조사/비양태어미/문법적 접미사]’와 같이 세울 수 있다. In this paper, we dealt with some problems such as distinction between free morpheme and bound morpheme; and distinction between lexical morpheme and grammatical morpheme. We accounted for these problems as follows. First, principles of morpheme identification are diverse because of purpose, subject, and methodology of studies. Second, the morpheme analysis of Korean is based on IA(Item-and-Arrangement) model. Identifying morpheme from word-form, researchers must consider that paradigmatic relation, syntagmatic relation, and phonological manifestation are important criteria. Third, empty morph(eme) and zero morpheme are useful concept in itself. However, it is highly desirable that these concepts do not be used at random. Fourth, empty morph and fossil are cross-classified into ‘empty morph and fossil’, ‘empty morph but not fossil’, ‘fossil but not empty morph’ in the light of correlation of fossilization with empty morphs. Fifth, dependency is classified into morphological dependency, syntactic dependency, and semantic dependency. In accordance with the realization of these dependencies, we can postulate a scale of the dependency of morphemes. Sixth, we also postulated provisionally the lexical hierarchy of Korean morphemes.

형태소의 유형

이정택(Jungtag Lee) 서울여자대학교 인문과학연구소 2010 인문논총 Vol.19 No.-

In this article, I tried to clarify the academic definitions of several morpheme types and the differences between counter types, namely free morpheme vs bound morpheme, root vs affix, full morpheme vs empty morpheme, lexical morpheme vs grammatical morpheme. The results of this study are as follows. 1. The meanings of 'free' and 'bound' in free morpheme and bound morpheme should be based on the meanings of free form and bound form in Bloomfield(1933). That is to say, free morpheme and bound morpheme should be defined as the morphemes which can be uttered in isolation and the ones which can not be uttered so respectively. 2. We cannot find any deference between root and full morpheme and between affix and empty morpheme according to their references. The terms 'root' and 'affix' were designed to focuss their roles in word formation. Then we don't have to use all of these terms, because of their references. I think just one couple of terms will be enough, namely full morpheme and empty morpheme or root and affix. 3. We cannot find any deference between the meanings of full morpheme and lexical morpheme and between the meanings of empty morpheme and grammatical morpheme. Even though the terms 'lexical morpheme' and 'grammatical morpheme' were made more recently, they can make some confusions. So we had better use the terms 'full morpheme' and 'empty morpheme'.

이합사와 이합사간의 품사 및 형태소 결합방식 연구 : 동소동서사를 중심으로

김진호 한국교통대학교 2020 한국교통대학교 논문집 Vol.55 No.-

The words with the same morpheme is words and words of the same morpheme in contact each other on the formation of the phenomenon. A group of words of the same morpheme, each call for the words with the same morpheme. The words with the same morpheme can be divided into two types; one is the same order of morpheme of the word (ex: 得力benefit from - 得⫽力capable), another is the combination of morpheme inversion of the words(ex: 考期date of an examination - 期考terminal examination). In the words with the same morpheme part of speech and some are the same, some are different. In the structure of some morphemes are the same, some are different. In this paper, ≪Modern Chinese dictionary≫ (1996) in the morpheme-same order words as the object, on the similarities and differences of the part of speech and the morpheme structure format in Separate-combinative compounds and Separate-combinative compounds research.

중국어 동소사의 품사와 형태소 결합방식 연구

김진호,현성준 한국중국문화학회 2017 中國學論叢 Vol.0 No.56

The words with the same morpheme is words and words of the same morpheme in contact each other on the formation of the phenomenon. A group of words of the same morpheme, each call for the words with the same morpheme. The words with the same morpheme can be divided into two types; one is the same as the sequence of morpheme word order (for example: 海口 - 海口), another is a combination of the turn down order word allotropy word (such as: 力量 - 量力). In twords with the same morpheme part of speech and some are the same, some are different. In the structure of some morphemes are the same, some are different. In this paper, 《Modern Chinese dictionary》 (1996) in the morpheme-inversed words as the object, on the similarities and differences of the part of speech and the morpheme structure format research.

單語와 離合詞간의 品詞 및 形態素 結合方式 硏究 -同素異序詞를 中心으로-

김진호 韓國交通大學校 2023 한국교통대학교 논문집 Vol.58 No.-

The words with the same morpheme is words and words of the same morpheme in contact each other on the formation of the phenomenon. A group of words of the same morpheme, each call for the words with the same morpheme. The words with the same morpheme can be divided into two types; one is the same order of morpheme of the word (ex: 通風 - 通⫽風), another is the combination of morpheme inversion of the words(ex: 喜報 - 報⫽喜). In the words with the same morpheme part of speech and some are the same, some are different. In the structure of some morphemes are the same, some are different. In this paper, ≪Modern Chinese dictionary≫ (1996) in the morpheme-inversed order words as the object, on the similarities and differences of the part of speech and the morpheme structure format research.

단어와 이합사간의 품사 및 형태소 결합방식 연구 : -동소동서사를 중심으로-

김진호 한국교통대학교 2019 한국교통대학교 논문집 Vol.54 No.-

The words with the same morpheme is words and words of the same morpheme in contact each other on the formation of the phenomenon. A group of words of the same morpheme, each call for the words with the same morpheme. The words with the same morpheme can be divided into two types; one is the same order of morpheme of the word (ex: 點心- 點⫽心), another is the combination of morpheme inversion of the words(ex: 利權- 權利). In the words with the same morpheme part of speech and some are the same, some are different. In the structure of some morphemes are the same, some are different. In this paper, ≪Modern Chinese dictionary≫ (1996) in the morpheme-same order words as the object, on the similarities and differences of the part of speech and the morpheme structure format research.

형태소의 유형 분류

최웅환(Choi, Ung-hwan) 한국문화융합학회 2008 문화와 융합 Vol.30 No.-

The purpose of this study is to review a classification of Morpheme in Korean. We did not discuss so that there was depth about classification of Morpheme in Korean. We are accepting concept and classification of Morpheme which was insisted by Bloomfield, L(1993). Classification of Morpheme was review in this study. We pay attention to three facts : First, Morpheme is minimal meaningful unit in the phonological word. Morpheme is categorized to Free Morpheme and Bound Morpheme. Second, Free Morpheme is not concerned at scope(phonological word) or keep unique meaning. Bound Morpheme keep unique meaning in scope(phonological word) only. Third, Lexical characteristic and grammatical characteristic is information of Morpheme only.

CNN을 적용한 한국어 상품평 감성분석

박현정(Hyun-jung Park),송민채(Min-chae Song),신경식(Kyung-shik Shin) 한국지능정보시스템학회 2018 지능정보연구 Vol.24 No.2

With the increasing importance of sentiment analysis to grasp the needs of customers and the public, various types of deep learning models have been actively applied to English texts. In the sentiment analysis of English texts by deep learning, natural language sentences included in training and test datasets are usually converted into sequences of word vectors before being entered into the deep learning models. In this case, word vectors generally refer to vector representations of words obtained through splitting a sentence by space characters. There are several ways to derive word vectors, one of which is Word2Vec used for producing the 300 dimensional Google word vectors from about 100 billion words of Google News data. They have been widely used in the studies of sentiment analysis of reviews from various fields such as restaurants, movies, laptops, cameras, etc. Unlike English, morpheme plays an essential role in sentiment analysis and sentence structure analysis in Korean, which is a typical agglutinative language with developed postpositions and endings. A morpheme can be defined as the smallest meaningful unit of a language, and a word consists of one or more morphemes. For example, for a word 예쁘고, the morphemes are 예쁘(= adjective) and 고(=connective ending). Reflecting the significance of Korean morphemes, it seems reasonable to adopt the morphemes as a basic unit in Korean sentiment analysis. Therefore, in this study, we use morpheme vector as an input to a deep learning model rather than word vector which is mainly used in English text. The morpheme vector refers to a vector representation for the morpheme and can be derived by applying an existent word vector derivation mechanism to the sentences divided into constituent morphemes. By the way, here come some questions as follows. What is the desirable range of POS(Part-Of-Speech) tags when deriving morpheme vectors for improving the classification accuracy of a deep learning model? Is it proper to apply a typical word vector model which primarily relies on the form of words to Korean with a high homonym ratio? Will the text preprocessing such as correcting spelling or spacing errors affect the classification accuracy, especially when drawing morpheme vectors from Korean product reviews with a lot of grammatical mistakes and variations? We seek to find empirical answers to these fundamental issues, which may be encountered first when applying various deep learning models to Korean texts. As a starting point, we summarized these issues as three central research questions as follows. First, which is better effective, to use morpheme vectors from grammatically correct texts of other domain than the analysis target, or to use morpheme vectors from considerably ungrammatical texts of the same domain, as the initial input of a deep learning model? Second, what is an appropriate morpheme vector derivation method for Korean regarding the range of POS tags, homonym, text preprocessing, minimum frequency? Third, can we get a satisfactory level of classification accuracy when applying deep learning to Korean sentiment analysis? As an approach to these research questions, we generate various types of morpheme vectors reflecting the research questions and then compare the classification accuracy through a non-static CNN(Convolutional Neural Network) model taking in the morpheme vectors. As for training and test datasets, Naver Shoppings 17,260 cosmetics product reviews are used. To derive morpheme vectors, we use data from the same domain as the target one and data from other domain; Naver shoppings about 2 million cosmetics product reviews and 520,000 Naver News data arguably corresponding to Google’s News data. The six primary sets of morpheme vectors constructed in this study differ in terms of the following three criteria. First, they come from two types of data source; Naver news of high grammatical correctness and Naver shopping’s cosme