An Analysis of Research Trends in Computer Security over the Last Decade

Abstract

Recently, text mining has popularly been used not only to identify important topics but also predict future tends in a research field. In this paper, we apply text mining to analyze the research trends in the field of computer security over the last decade. We collected 2,256 papers published in the four major security conferences (CCS, NDSS, USENIX Security and S&P) between 2008 and 2017. We extracted keywords from the title and abstract in each paper and used them to identify important topics based on word frequency counts for each keyword. We analyzed the rankings and the changes in the most frequent keywords every year. Our observations from the analysis would be useful for understanding the popularity of computer security research, and predict potentially important research topics (e.g., SGX, Fuzzing, Blockchain, IoT and SDN) in the near future.