논문 상세보기
Using Textual Analysis of News Articles to Identify State’s Nuclear Activity
Monitoring a state that intentionally hides its nuclear activity via open-source information is akin to looking through a black box. Direct information on the state’s nuclear activity remains in the dark, leaving scholars to speculate how much nuclear material or warheads are being produced. Nevertheless, a state’s nuclear program consists of a complex network that ranges from producing weapon-grade nuclear materials by operating its nuclear facilities to securing resources to fund these activities. These indirect activities allow a narrow window of opportunity for researchers to map a state’s activity that sometimes may not be directly linked to nuclear activity per se but is significant to maintaining and operating its nuclear program. These may include malicious cyberattacks to steal or launder cryptocurrency and facilitating cooperation with fellow rogue states that do not comply with the NPT and nuclear nonproliferation regime. The problem lies in how researchers can map this network. Much of the literature that uses text analysis uses data from either (1) formal statement, reports, and documents or (2) journal articles to extract relations between topics that is otherwise difficult to surmise. This study, however, analyzes news articles containing keywords related to a states’ nuclear activity such as international sanctions, trade activities, other states’ policy etc. While news articles fail to live up to the academic rigor of journal articles and unlike formal documents may sometime contain misinformation or incorrect facts, they are a valuable medium to show the day-to-day activity of a state. Although bias may exist as particular news articles may or may not be chosen for text analysis, by using articles collected from 2021 to 2022, this study argues it is enough data to show a short-term trend in nuclear activity.
