Exploring Knowledge Graphs with Directed/Mixed Graph Isomorphism

Recent years have witnessed a rapid growth of knowledge applied extensively in both academia and industry. This thesis mainly focuses on exploring KGs with graph-based computational approaches, including a novel algorithm for directed/mixed graph isomorphism in polynomial time. Furthermore, the effectiveness and efficiency of the proposed approaches are verified by their application on the identification of money laundering behavior in transactional KGs and the construction and exploration of viral/chemical similarity KGs regarding novel coronavirus disease 2019 (COVID-19). The proposed approaches in this thesis may contribute to more informed decision-making in a wide range of areas.