Inferring user goals from sets of independent queries in a multidatabase environment

Abstract

Loosely coupled data integration among networked sources has become so ubiquitous over the recent years that many of the services and applications used daily are actually not monolithic information systems but rather collections of sources tied together. Instead of building centralized and large data sources (i.e., the Extract- Transform-Load method), many organizations and individuals are opting for a virtual database approach. Especially along with the advent of service-oriented architectures, it has become very easy to leave data in its original source and to instead recruit the service provided by that source as needed. This structure is seen in a variety of scenarios such as hybrid web applications (mash-ups), enterprise information integration models, aggregation services and federated information retrieval systems. Furthermore, individual users are often forced to procure and assemble the information they need from sources distributed across a network. © 2010 Springer-Verlag Berlin Heidelberg.