텍스트 마이닝 기법을 활용한 시조 분석 ―코로나-19 창작 시조를 중심으로

초록

Recently, humanity has entered a new era termed the fourth industrial revolution. Accordingly, the boundary between natural science and humanities is gradually fading, as research converges between the two disciplines. Research to analyze text in humanities using digital techniques for natural science has also been attempted in various ways.Thus,thispaperispartofabodyofworkaimingtoconfirmthepossibilityand effectiveness of sijo analysis using text mining techniques. Sijo is a shortened term for “songs of the time.” A kind of popular song that reflects the time, sijo has mirrored current times even in theCOVID-19pandemic,anissuethathasaffectedtheworldinrecentyears.Thiscouldalsobe confirmed through the modern trend that has recently been created. It is expected that the characteristics of modern sijo will be seen in the sijo created by students. Therefore, in this study, 195 sijo created by 192 students who took Korean Language and Literature major classes in 2021 and 2022 under the sanctions of "COVID-19" were analyzed. Through these sijo, we identified students' perceptions of COVID-19. The students' original poems were analyzed by applying all three methods of text mining techniques, and for the first time, sijo morphemes were analyzed using KoNLP, a Korean text analyzer to determine the frequency of word appearance in each work and embed it usingWord2Vec's function. Common words in the sijo works created in 2021and 2022 included "people," "we," "face," and "when," in addition to"COVID-19" and "mask." These words were identified as expressions of change, crisis, and willingness to overcome them. Next, we conducted an emotional word analysis through the KNU Korean Emotional Dictionary. The use of words containing strong emotional expressions in works created in 2022 tended to decrease from those in 2021, which may be due to the gradual easing of emotions as people adjust to the prolonged COVID-19 situation. In addition, we compared prominent appraisers in 2021 and 2022, and found that students' perceptions of COVID-19 are changing over time. Finally, the titles of the sijo were visualized as a word cloud and analyzed. Positive words tended to more frequently appear in titles in 2022 than in 2021, a conclusion similar to the findings of our emotional word analysis, so it can give its own meaning.

키워드