Research on quality evaluation and risk control methods of open source cybersecurity datasets

Open-source data is an important source of information for network threat detection, network threat intelligence, network risk management, and network security situational awareness. In the broad field of intelligence processing, open-source intelligence accounts for over $95 \%$ of all intelligence. Network security activities, such as intrusion detection, attack surface management, and incident response, are increasingly relying on open-source threat intelligence published by global security organizations and professional security companies. However, the quality parameters of opensource data, such as authority, expertise, domain specificity, real-time accuracy, and consistency, vary greatly. Low-quality open-source data can reduce the accuracy of analysis and impact subsequent decision-making and actions. To address the challenges of identifying and controlling the quality of open-source data, a data quality assessment system is proposed based on the characteristics of open-source data. It establishes comprehensive and practical dimensions, evaluation theories, and tracking methods for open-source data quality, supporting the enhancement of timeliness, integrity, and relevance of open-source data through on-net and off-net control measures. This helps unleash the value of data, ensuring timely and accurate analysis, traceability, and response to network security incidents.