딥러닝 기법을 활용한 북한의 대외인식 분석 연구 - 『로동신문』(2016-2019) 빅데이터를 중심으로 - = A deep learning-based analysis of awareness changes in North Korea : Focusing on Rodong Sinmun Big Datas from 2016 to 2019

다국어 초록 (Multilingual Abstract)

This article analyzed the changes in North Korea’s internal and external perceptions using the deep learning-based analysis about Rodong Sinmun from 2016 to 2019. Even if the growing peace mood that had stayed for a while on the Korean Peninsula now disappeared, the change in North Korea’s attitude over a short period of time was very unusual. It was attempted to provide a more detailed understanding about the possibility of future changes in North Korea by tracking changes in North Korea’s perceptions of South Korea and its perceptions of peace and threats. As a result, the main threat was the US for North Korea’s perspective, and it was also confirmed that South Korea could be perceived as a threat by insulting on the dignity of the supreme leader-Kim Jongeun. The fact whether South Korea’s governments consist of progressive or conservative party doesn’t seem to have significant impacts on North Korea’s perception. Expectations for improvement in inter-Korean relations in 2018 are highly likely to have been perceived by North Korea as an escape from the crisis through the ease of economic sanctions. The reason is that North Korea’s perception about the threat has returned to the same level as before after the severance of US-DPRK dialogue, and it could be revealed by ordering slogans for domestic economic development. This study applied the machine learning technique to the full text of the Rodong Sinmun for 4 years, confirming the changes in the main vocabulary of the Rodong Sinmun.