In an uncooperative distributed retrieval environment, query based sampling is required to probe selected sources in order to obtain an underlying representations. As part of this research more appropriate methods for the evaluation of a resource description were developed before proposing adaptive query based sampling (AQBS). AQBS focused on sampling a particular sub-sample of the source which was representative of the user’s needs. The idea was that this would provide a better representation w.r.t the users, as opposed to relying on an unbiased and potentially irrelevant representation.

See Online Publications For More Details

This work was performed in conjunction with Mark Baillie and Fabio Crestani and under the remit of two european projects: PENG and DILIGENT.