Mining service integration opportunities towards joined-up government

Service integration is central to joined-up government initiatives and requires information on the collaborators and the services they offer, roles of different actors, the resources required, and their goals (individual and shared). These information are largely available in unstructured forms on government portals, publications and other textural sources. This paper explores semantic text mining for extracting service-related information from such sources using Natural Language Processing techniques supported by Service-Oriented Process Ontologies. Our solution framework consists of the following steps: (1) creating domain and service-oriented process ontology, (2) extracting servicerelated information from textual sources based on the ontology, and finally (3) mining relationship among the services based on the extracted information in Step 2 linked with a pre-defined hierarchy of service delivery goals specifying the objective(s) to be achieved among the orchestrated services. We describe our approach to these tasks and discuss the progress of the work, our experiences and the challenges encountered so far.