빅데이터 적정 텍스트 추출을 위한 언어학적 접근-‘학교폭력’ 관련 텍스트를 중심으로- (Linguistic approach for Document Classification on Big Data - focused on ‘School Violence’ texts-)

서지정보

· 발행기관 : 한국텍스트언어학회
· 수록지 정보 : 텍스트언어학 / 36권 / 149 ~ 177페이지
· 저자명 : 남길임, 송현주, 이수진

초록

Discussions of the text types or the issuesabout Korean language processes in the field of big Data studies havebeen unnoticed. This study aims to introduce the methodology of theexclusively collecting the ‘school violence experience’ texts in orderto prevent school violence and to catch some of the signs by analyzingthe unstructured data created in real time on a web. A certain texttype has explicit markers revealing some aspects of the topics or someinformative functions of texts. These markers are able to be appliedto extraction of appropriate texts. Moreover, individual words andmorphemes can be classified into some semantic categories and afrequency of those categories becomes a factor which contributes todistinguish a particular text from the other. This study assumes thatthere are keywords which occur with significantly high frequency atthe level of word or morpheme, and semantic categories which appearsignificance in the level of meaning in our ‘school violence experience’texts, so that these keywords and semantic categories can be usedfor the classification of appropriate texts according to topics.
Conclusively, the goal of this study is to examine the way of theclassification of appropriate texts by those linguistic features.

영어초록

Discussions of the text types or the issuesabout Korean language processes in the field of big Data studies havebeen unnoticed. This study aims to introduce the methodology of theexclusively collecting the ‘school violence experience’ texts in orderto prevent school violence and to catch some of the signs by analyzingthe unstructured data created in real time on a web. A certain texttype has explicit markers revealing some aspects of the topics or someinformative functions of texts. These markers are able to be appliedto extraction of appropriate texts. Moreover, individual words andmorphemes can be classified into some semantic categories and afrequency of those categories becomes a factor which contributes todistinguish a particular text from the other. This study assumes thatthere are keywords which occur with significantly high frequency atthe level of word or morpheme, and semantic categories which appearsignificance in the level of meaning in our ‘school violence experience’texts, so that these keywords and semantic categories can be usedfor the classification of appropriate texts according to topics.
Conclusively, the goal of this study is to examine the way of theclassification of appropriate texts by those linguistic features.

참고자료

· 없음