Towards a Generic Knowledge Graph Construction Framework for Privacy Awareness

Christina Karalka, Georgios Meditskos, Maria Papoutsoglou, Nick Bassiliades

Abstract: Knowledge graphs (KGs) organize data from multi-ple sources, capturing information about entities of interest in a given domain or task, such as people, places, or events, and forge connections between them. In this paper, we introduce a generic framework for building knowledge graphs designed to enhance data privacy through semantic interpretation. We demonstrate the effectiveness of our framework by applying it to the healthcare sector, where it helps organize and analyze com-plex information, support data analysis, improve decision-making processes, and uncover hidden relationships between entities. Our approach leverages domain-specific ontologies like SNOMED CT and integrates vector databases to assess and mitigate privacy risks. By using semantic techniques, we enhance the robustness of data against reidentification attacks and suggest appropriate de-identification methods. The integration of SNOMED with vector databases allows for efficient storage, retrieval, and analysis of high-dimensional healthcare data, facilitating advanced data an-alytics and knowledge discovery while maintaining data privacy. Through this framework, we aim to provide sufficient insights for identifying privacy vulnerabilities and ensuring the security and usability of sensitive health information.