Typed Multirelational Attributed Metagraphs for Modeling Criminal Events and Hidden Connections
摘要
A formal approach to the representation and analysis of criminal events based on a typed multirelational attributed metagraph is proposed. Events are treated as structured objects with internal organization, allowing for the integration of event-based and object-oriented descriptions within a single mathematical model. The concepts of typed adjacency and incidence are introduced, including their “weak” and “strong” variants for subsets of relation types, along with their corresponding matrix representations. Typed vertex degrees are defined, enabling the assessment of the role and significance of entities across various semantic contexts. The introduction of typed paths facilitates the formalization of semantically controlled chains of connections, providing a foundation for identifying indirect and potentially hidden interactions between events and individuals.