OBDAIR: Ontology-Based Distributed framework for Accessing, Integrating and Reasoning with data in disparate data sources

Abstract

Abstract The correlated exploitation of disparate and heterogeneous data sources is important to the efficacy of many analytics tasks. Currently in application domains of major interest, such as in the maritime and aviation domains, available technology provides real time surveillance data from moving entities, which together with archival static data, can be processed in an integrated way to detect complex events and support decision making. The variety of data in disparate sources, the heterogeneity of data formats, as well as the volume of data, make data retrieval, integration, and especially reasoning with these data, challenging tasks. This paper presents an ontology-based distributed framework that addresses conjunctively these challenges: Data retrieval, integration and reasoning with data from heterogeneous static or regularly updated data sources. The proposed OBDAIR framework provides the means to support building scalable data-driven domain-specific applications that support decision-making and problem-solving. This is achieved by processing large volumes of heterogeneous data close to the sources, supporting knowledge generation in a distributed/decentralized but still unified manner. OBDAIR integrates modular ontology representation frameworks and ontology-based data access frameworks: This article presents an instantiation of OBDAIR using the modular ontology representation framework E − SHIQ , and the Ontop ontology-based access system. This OBDAIR instance has been evaluated at recognising important complex events in the maritime domain using real-world data. Experiments show the potential of OBDAIR to detect complex events in large geographic areas with computational efficiency.