Ontology Driven Securities Data Management and Analysis

With the increase of fraudulent transactions in world wide securities market, it is critical for regulators, investors and public to accurately find such business practices to avoid serious loss. This paper makes a novel attempt to efficiently manage securities data and effectively analyze suspicious illegal transactions using an ontology driven approach. Ontology is a shared, formal, explicit and common understanding of a domain that can be unambiguously communicated between human and applications. Here, we propose an ontology model to characterize entities and their relationships in securities domain based on a large number of case studies and industry standards. Securities data, (namely said financial disclosure data, such as annual reports of listed companies), are currently represented in XBRL format and distributed in physically different systems. These data from different sources are firstly collected, populated as the instances of the constructed ontology and stored into an ontology repository. Then, inference is performed to make the relationships between entities explicit for further analysis. Finally, users can pose semantic SPARQL queries on the data to find suspicious business transactions following formal analysis steps. Experiments and analysis on real cases show that the proposed method is highly effective for securities data management and analysis.