Ontologies for Agent-Based Information Retrieval and Sequence Mining

In this paper, we present two very practical problems in the areas of distributed information retrieval and pattern mining, as well as our proposed solutions via the use of intelligent agents and domain ontologies. The first problem is to retrieve data from heterogeneous distributed data sources with a specific application to distributed Earth Science data archives. Our proposed approach is to develop an engine which acts as an interface agent by presenting users with the appearance of a single, unified, homogenous data source based on a domain ontology of Earth Science terminology. Users can then pose high-level declarative queries against this view. The system then translates each query into a set of sub-queries and spawns mobile agents to retrieve data corresponding to each sub-query. The second problem is to predict significant world events at multiple levels of abstraction by analyzing a collection of events over a period of time in order to generate sequential patterns. We specifically focus on predicting terrorist actions by analyzing terrorist group activities over time. We employ a hierarchical taxonomic organization of contextual event types to obtain higher-level abstractions of observed low-level events. With this approach, significant events can be predicted at multiple levels of abstractions with associated confidences. Although we have addressed these two problems by building prototypes in two different domains, their combination offers a powerful agent-based tool that can assist scientists and analysts by automatically retrieving and mining data collected from multiple distributed data sources. Thus with the use of relevant domain ontologies, the problems of data retrieval and pattern discovery can be combined and automated in a single, elegant system.