자연어처리를 이용한 미국의 코로나 이후 증오범죄 변화분석

Abstract

A total of eight people died in a shooting spree in Atlanta, Georgia, on March 16, 2021, and six of the deaths were Korean women. This has caused public anger not only in Korea but also in other Asian countries. Like this incident, indiscriminate attacks accompanied by racism have been increasing rapidly since COVID-19. The purpose of this study is to identify the relationship between race-related hate attack types and COVID-19 in the United States through unstructured text data analysis and discuss countermeasures. From January 2018 to March 2020, articles on hate crimes in the United States (N=485) posted on famous US news sites such as the Washington Post, the LA Times, and the New York Times were collected through webcrawling and the process of refining was conducted to verify equity to use high-quality articles for analysis. Simultaneous Appearance Word Network (TF-IDF) analysis and LDA (Latent Dirichlet Allocation Modeling) model were applied to classify meaningful topics. As a result, there was a difference in major topics before and after COVID-19. This research method finally discussed the policy implications and limitations that contribute to the establishment of hate crime prevention measures for Asians.