Many scientific experiments produce large amounts of data using high-throughput devices. Knowledge Discovery systems are used in order to analyse this type of data. However, generic laboratory systems do not provide any contextual information about the system that is being studied. In these situations, Knowledge Discovery can be aided and validated by the use of Information integration tools. In this paper, we introduce InfoGrid, a data integration middleware engine, designed to operate under a Grid framework. It focuses on providing information access services and offers all users a query system which is able to retain the familiarity with their specific scientific applications while being diverse, flexible and open at the same time. The assumption here is that defining a common language for all queries is not desirable.Using this design, we show how the InfoGrid architecture can be used to provide contextual features for a data table to be used for analysis (i.e. the Annotation Problem). We also show how it can be used to find relevant background knowledge for a user (i.e. the Information Comprehension Problem). Both of these issues are very often encountered in Knowledge Discovery tasks, which we illustrate by means of a real-world example.
pubs.doc.ic.ac.uk: built & maintained by Ashok Argent-Katwala.