텍스트 마이닝을 통한 치안 수요 자동범주화 연구

Abstract

In order to enhance the perception of public safety and security satisfaction level, the government should grasp the exact demand for police activities and public security. In some areas, open-ended surveys were conducted to figure out the exact demand for police activity, but there were limitations in terms of the time and cost involved in categorization and data analysis. Therefore, this study proposed an automatic categorization technique using text-mining as one of the ways to analyze open-ended responses. We collected response data from 480 residents in the A-area and compared the categories derived from both subjective categorization and automatic categorization. The majority of the major categories with high frequency of emergence were extracted in common from both methods. So the possibility of using automatic categorization technique was confirmed. Although it was confirmed that some demands that could be omitted from subjective categorization can be captured through automatic categorization, the limitations were also identified that it is difficult to extract categories that are composed abstractly or comprehensively. However, despite these limitations, the possibility of the use of automatic categorization technique in the survey of security needs was confirmed, given that text-mining technique can be used to summarize the topics in the data and derive some criteria for the composition of the categories.