차별적인 보안 활동 지원을 위한 데이터 가치 분석 연구

초록

As data has become a core asset in economic activities in the era of the data economy, the financial industry, which handles highly sensitive information such as personal data, necessitates differentiated security systems. However, traditional classification processes often rely on subjective human judgment, resulting in inconsistent and unreliable evaluations that fail to fully capture the value of the data. This study proposes an objective model for analyzing the value of financial documents and classifying their security levels. Using the Delphi method, criteria for security classification were established, and the validity of data labeling was verified by comparing expert opinions with classification results from GPT-3.5-Turbo. Subsequently, an automated security classification model was developed using the ALBERT model, categorizing financial data into three levels. This approach enables an objective and systematic framework for data management and security in the financial sector. It allows organizations to implement tailored security strategies based on data sensitivity and utility, ultimately mitigating the risk of information breaches.

키워드