『동광』의 언어적 근대성 -데이터과학적 분석-

조은경 ( Jo¸ Eunkyoung ),드프레메리웨인 ( De Fremery¸ Wayne ) 한국언어문화교육학회 2021 언어와 문화 Vol.17 No.3

This study investigates Korean linguistic modernity as it is expressed by the colonial-era periodical 『Donggwang』 (Eastern Light). To compliment previous qualitative research on Korean linguistic modernity, we present a quantitative analysis of textual data associated with 『Donggwang』 made available by the National Institute of Korean History. Rather than working deductively to test any particular theory of linguistic modernity, we work inductively to describe language use in an important modern periodical. Assessing the mixed script employed in 『Donggwang』 with quantitative methods, we discern an increase in the use of Sino-Korean over the course of the periodical’s publication in the 1920s and 1930s. Information-theoretic measures of mutual information substantiate some previous claims about authors appearing in 『Donggwang』. They confirm, for example, the unsurprising fact that Kim Yungyeong is strongly associated with the study of the Korean language. They also suggest new ways to understand the disposition of Kim’s writing. For example, our analysis shows that Kim’s texts can be more easily associated with words such as gugeo (national language), gukeohak (national linguistics), gukga (nation) than words such as Joseon (Joseon Korea) or Joseoneo (Joseon language). Visualizing word embeddings and related terms substantiate previous scholarship. We confirm the term hakkyo (school) often denotes modern educational institutions, for example, and the terms munja (written characters), yeongu (research) and gwahak (science) are frequently associated. We also identify new avenues for investigation. For example, our analysis suggests the terms such as jabonjuui (capitalism), segye (world) and gonghwang (panic) are important to authors appearing in 『Donggwang』. In sum, we argue that data scientific techniques like those we employ can enrich and extend qualitative formulations of concepts such as Korean linguistic modernity. (Sogang University)

정보 기술 분야에서 개체명 동의어 연구

조은경(Eunkyoung Jo) 언어과학회 2014 언어과학연구 Vol.0 No.69

Named Entities are very important linguistic elements in information technology domain which carries out natural language processing. Specifically, This paper shows why problematic the synonyms of named entities in linguistic information processing application areas like search engine and dialog system are. We investigate how the synonyms of named entity are used based on the corpus of web document and then propose lexicon improving method to facilitate the method of imformation processing about named entity synonyms in various nlp applications.

대화 시스템에서의 조응어 해석

조은경(Eunkyoung Jo),서정연(Jungyun Seo) 한국정보과학회 언어공학연구회 2004 한국정보과학회 언어공학연구회 학술발표 논문집 Vol.16 No.1

조응어 해석(Anaphora Resolution)은 여러 응용 분야에 적절히 적용됨으로써 자연어 이해에서의 양적, 질적 향상을 가져 올 수 있다. 구어에서 조응어(anaphor)는 화자와 청자간에 공유된 정보를 간략하게 나타내는 형태이다. 특히, 구어를 주된 사용 언어로 하는 대화 시스템에서의 조응어의 양상을 살피고 해석하는 문제는 인간과 기계간의 의사소통(man-machine communication)을 온전히 하는 과제이다. 이 논문에서는 인간과 기계간의 대화 시스템(man-machine dialogue system)에서 흔히 쓰일 수 있는 조응 표현에 중점을 두어 그 유형을 살피고 해석의 방식을 제시하도록 한다.

시간 부사와 시제 표지의 공기에 관한 말뭉치 기반 시간 지향성 해석

조은경(Jo, Eunkyoung) 담화·인지언어학회 2016 담화와 인지 Vol.23 No.1

This paper looks for the quantitative tendency oriented to temporal categories of past or future formed by the co-occurrence of tense markers and time adverbs which are the features of representing temporal concept and investigates how tense markers and time adverbs are juggled to represent temporal concept of affairs. Most previous studies deal with the constraints and tendencies of using time adverbs and tense markers and then focus on the distinction between tense and aspect as grammatical categories. However, we do not focus on the distinction, but on the tendency of temporal concept of past or future categories formed by the co-occurrence of tense markers and time adverbs and how tense markers and time adverbs are juggled to represent temporal concept of affairs. For this, we use a quantitative analysis based on ‘Sejong written pos-tagged corpus’. We extracts the concordances of which time adverbs and tense markers co-occur in a clause. As for quantitative methods, we use co-occurrence frequency and relative frequency. As a corpus-based approach, we looked for the temporal orientednesses which the time adverbs given tense markers represent and the tense markers given time adverbs represent respectively. Also, we investigate how a future-oriented time adverb co-occurring with a past-tense marker plays a role in temporal concept of context and a past-oriented time adverb co-occurring with a future-tense marker does. Furthermore, we investigate how tense markers co-occurring with different time adverbs represent their temporal concepts in context.

『별건곤』 의 텍스트 근대성 : 데이터과학적 언어분석

조은경(Jo Eunkyoung),드프레메리웨인(Wayne de Fremery) 한국문화융합학회 2021 문화와 융합 Vol.43 No.10

이 연구는 1926-1934년의 간행물인 근대잡지 별건곤 이 텍스트 자료로서 가진 언어적 근대적 양상을 분석한다. 국사편찬위원회에서 제공하는 별건곤 의 기사를 모두 수집하여 데이터과학적, 언어분석적 방법을 적용하였다. 1930년대에는 논술 기사와 문학 기사가 감소한 반면, 광고 기사(소식)와 잡저는 증가하였고, 한문은 1930년대에덜 널리 사용됨을 알 수 있었다. 이것은 편집자들이 광범위한 대중적 독자층을 참여시키려는 열망을 나타낸다. DF(Document Frequency)를 사용하여, [별건곤]에 대한 이전 문헌 연구에서는 비교적 중요하게 다뤄지지 못한근대적 특성을 나타내는 단어 목록을 추출할 수 있었다. ‘비누’, ‘위스키’ 등 위생 및 문화상품 관련 단어, ‘피임’,‘로만수’(로맨스) 등 가족 및 관계 관련 단어 포함 ‘인쇄물’, ‘중앙일보’와 같은 근현대적 매체 특성을 뜻하는 단어,근대화 시기 글로벌 환경을 드러내는 ‘아라비아’, ‘외국어’와 같은 단어를 적정한 분포로 찾아내었다. 또 단어 임베딩을 시각화하여 ‘여성’과 ‘결혼’과 같은 이전 연구자들이 [별건곤]에 관해 많이 논의한 주제어뿐만 아니라 이전 연구에서 간과된 많은 개념 단어를 드러낸다. ‘도시화’와 ‘산업’, 그리고 ‘노동’과 ‘노동자’ 등이 그러하다. 이러한 단어들은 별건곤 을 통해 식민지 시기 근대화 과정의 특성을 이해하는 새로운 관점을 제안한다. 또한 별건곤 의 근대적 특성의 단어들이 현대어 언어 모델을 기반으로 하여서도 여전히 예측 가능한 개념적 단어임을 보여준다. This study investigates Korean textual modernity as expressed by the colonial-era periodical[Byeolgeongon]. To compliment previous qualitative research associated with the periodical, wepresent a quantitative analysis of textual data associated with [Byeolgeongon] made available bythe National Institute of Korean History. We discern that over the run of [Byeolgeongon] between1926 and 1934 the number of full-length articles (nonsul) and literary works decrease while gossip(sosik) and miscellaneous articles (jabjeo) increase. The use of Sino-Korean also decreases overthe run of the journal. We hypothesize that these patterns suggest a desire on the part of editorsto engage a broad popular readership. Using DF (Document Frequency), we retrieve a list of wordsthat helps to reveal aspects of Korean textual modernity overlooked by previous research on[Byeolgeongon]. Our analysis suggests the importance of words associated with hygiene andcultural products in [Byeolgeongon], such as ‘binu’(soap) and ‘wisuki’(whiskey). Words associatedwith family and relationships, such as ‘piim’(contraception) and ‘romansu(romance), media products,such as ‘inswaemul’(printed materials) and ‘Jungang ilbo’(Jungang Daily News), and globalperspectives, such as ‘Arabia’(Arabia) and ‘oegukeo’(foreign language), also formulate Korean textualmodernity as expressed by [Byeolgeongon]. Visualizing word embeddings, we discover relationshipsthat previous scholars have discussed, such as associations between ‘yeoseong’(woman) and‘gyeolhon’(marriage), but also many relationships that previous scholars have overlooked in thecontext of [Byeolgeongon], such as ‘dosihwa’(urbanization) and ‘saneop’(industry), as well as‘nodong’(labor) and ‘nodongja’(laborers). These relationships suggest new avenues for understandingKorean textual modernity as expressed by [Byeolgeongon]. Additionally, we confirm that thesewords can be inferred using Korean Language Understanding Evaluation models, suggesting thatthe words can also help to organize our conceptual understanding of textual modernity in Korea

역사 자료 어휘 분석기를 위한 어휘부 개발 방법

조은경(Jo, Eunkyoung),한영균(Han, YoungGyun) 한국사전학회 2016 한국사전학 Vol.- No.27

This paper explores some methods and techniques for structuring and acquiring lexical resources from corpus and diachronic knowledge for lexical analyser of historical text. Firstly, we consider the lexicon as a component of lexical analyser and playing a key role in the analyser. Secondly, we see the problem of correcting the annotated historical text corpus. Thirdly, we investigate and design the lexicon with affix features and stem features which are required to analyze words into morphological units. Besides, in this phase we consider Korean diachronic information and corpus quantitative information. Fourthly, we consider the methods for acquiring and estimating variable spellings of a lexical token. Finally, we can look forward to being able to take the lexicon with diachronic information as the resource for secondly created historical digital contents. All in all, this paper shows the methods for acquiring lexicon from data with quantitative and qualitative ways and the usage as a digital content with this lexicon.

결정 트리를 이용한 지시 표현 ‘것’의 구별

조은경(Eunkyoung Jo),김학수(Harksoo Kim),서정연(Jungyun Seo) 한국정보과학회 2007 정보과학회논문지 : 소프트웨어 및 응용 Vol.34 No.9

지시 표현 ‘것’은 한국어 대화에서 자주 등장하는 표현이지만, 그 자체로서 대명사나 한정 명사구와 같은 지시 표현이 아니고, 비지시적인 표현으로 쓰이는 ‘것’과 구별되지 못했기 때문에 지시 해석(reference resolution)에 관한 기존 연구에서 제대로 다루어지지 못했다. 이러한 문제를 해결하기 위해 ‘것’이 가지고 있는 언어학적 속성과 담화 상의 속성을 기반으로 하여 자질 집합을 설정하고, 결정트리를 이용하여 ‘것’을 구별하는 방법을 제안한다. 이 방법에 의한 시스템은 비지시 표현의 것에 대해 92%, 지시 표현의 것에 대해 82%의 F-measure를 보였으며, 전체적인 분류 성능은 89%였다. 이는 패턴에 따른 규칙을 적용한 분류 성능에 비해 약 15% 가량 향상된 결과이다. Referential expression ‘Geot’ is often occurred in Korean dialogues. However, it has not been properly dealt with by the previous researchers of reference resolution, since it is not by itself the referential expression like pronoun and definite noun phrases, and it has never been discriminated from non-referring ‘geot’. To resolve this problem, we establish a feature set which is based on the linguistic property of ‘geot’ and the discourse property of its text, and propose a method to identify referential ‘geot’ from non-referring ‘geot’ using decision tree. In the experiment, our system achieved the F-measures of 92.3% for non-referring geot and of 82.2% for referential geot and the total classification performance of 89.27%, and outperformed the classification system based on pattern rules.

책 리뷰 말뭉치를 활용한 형용사, 부사 감성 사전 개발

이상진(Lee, Sangjin),조은경(Eunkyoung Jo) 언어과학회 2020 언어과학연구 Vol.0 No.92

This paper presents a method of lexicon building for sentiment analysis using book review data. We show how to build sentiment lexicons automatically using book review data that includes readers" evaluation scores, basic positive and negative sentiment word lists using a semantic orientation measure, and sentiment word lists using Word2Vec. Specifically, we show how to automatically generate a useful lexicon of adjectives and adverbs for sentiment analysis. In summary, the process we describe is as follows: Basic word lists of positive and negative sentiments are created by relating word frequencies with reader"s book review scores. The expanded word lists are then built using the vector space representation of words and their cosine distance. This is made possible by assuming that words of similar sentiment and semantically co-related words will tend to appear together in the same context, and that they are represented proximately in the vector space.

안내 서비스 로봇의 형태 비례에 관한 연구

홍성수(Seongsoo Hong),조은경(Eunkyoung Jo),이예윤(Yeyoon Lee),한고은(Goeun Han),정정필(Jungpil Jung),박선민(Sunmin Park) 한국디자인학회 2020 한국디자인학회 학술발표대회 논문집 Vol.2020 No.11