An investigation of techniques for integrating Web-accessible data

The proliferation of World Wide Web accessible information sources has introduced the problem of integrating these sources to create unified views over them. Web-accessible data of interest can be customized and tied together in any number of ways and using a variety of possible techniques to result in structures known as datawebs. Collaborators and others with a desire for unified views of multiple datawebs face the challenge of being able to integrate such webs without actually exercising any control or ownership over them.; This thesis explores alternatives for organizing Web-accessible data in ways that make the automated integration of such data possible. We investigate techniques for integrating Web-accessible data and show that datawebs constructed following a uniform architecture enable a simplified approach to the design and development of lightweight, automated integration tools and techniques. Using our approach, we built the Web Integration Tool (WIT).; The goal of the tool is to simplify, and to the extent possible, automate the process of selectively integrating multiple datawebs into a unified dataweb. The resulting structures can then be accessed and efficiently navigated by collaborators and others using Web resources (and relationships among them) from the combined datawebs. WIT is also capable of mapping integrated results into an alternative, XML-based architecture.; This thesis contributes a method for automatically integrating Web-based resources from a wide variety of information spaces. The method enables a high degree of user selectivity with respect to the source and type of resources to be integrated. Using this method, the integration result can be further integrated with other Web-based resources, enabling arbitrary numbers of datawebs to be integrated. Our integration tool, WIT, implements our method to enable a fully automated approach to the integration of Web-based resources.