딥러닝을 활용한 시조의 유형 고찰 - 영남가단과 호남가단의 시조를 중심으로

초록

Artificial intelligence is rapidly developing day by day. Now it is becoming a part of our lives, not just business sites or engineers’ labs. The role of artificial intelligence is becoming increasingly important in literary research, and the necessity and possibility of artificial intelligence is increasingly proven in fields that target a variety of writers and works created over a long period of time, such as the founder. And artificial intelligence, ‘classification’ is often used regardless of life and research. Classification is used in many parts of everyday life, and classification is a very important methodology in literary research. Thus, this study attempted to analyze the Sijo by combining deep learning of artificial intelligence with the methodology of classification. Using a deep learning classification algorithm, 299 and 290 works of Yeongnam and Honam were trained, and 4,736 works of the “Korean Sijo Dictionary” were classified in earnest. As a result, 3,758 works (79.4%) were closely related to the two bands. This shows the status of the two bands in the history of Sijo literature. According to deep learning analysis of 299 works by artists from Yeongnam Gadan and 290 works by artists from Honam Gadan, 281 works for the former and 229 works for the latter were highly related to each band. This confirmed the potential for deep learning to be utilized in the Sijo study. On the other hand, although the frequency was not high, the 7th number of Yeongnam Gadan and 35th number of Honam Gadan were the opposite. This reflects differences in the relationship and creative environment of authors not included in the deep learning classification algorithm, which is a meaningful conclusion in that it suggests that external factors as well as internal characteristics of the work should be considered. Other writers who did not belong to any of the two bands were also able to be understood as a result of a combination of the writer’s life, friendship, and entering and leaving office. This study still remains level of attempt that examines classical literature as a new methodology called artificial intelligence. We will make up for this in the future.

키워드