A Review of Text-Based Information Security Rating: Fundamental Concepts, Methods, Datasets, Challenges, and Future Works

초록

The growing importance of critical data for national competitiveness and corporate survival underscores the need to classify and protect sensitive documents. Many researchers have introduced information security rating, a data classification method considering document security levels. However, research in this area has faced challenges in methodology development and application due to the lack of a standardized definition and the scarcity of survey papers exploring the latest research trends. To address these issues, this research proposes a standardized term, "information security rating," and establishes a systematic taxonomy of text-based information assets, including domain scope, methodology, and metrics. The primary contribution of this study is to comprehensively review the overall research trends, covering both administrative and technical methodologies, from rule-based methods to deep learning models. It also introduces representative datasets and various evaluation metrics, such as the CIA triad, impact factors, and text classification metrics. Furthermore, this study identifies and proposes five novel limitations from different perspectives, including the challenge of unbalanced confidential data, the need for alternative security evaluation metrics, and convergence approaches. Overall, this study will serve as a fundamental guideline, by providing insights into future research directions.

키워드