A wrapper&mediator prototype for web data warehouses

Abstract

There is a lot of information published on the Web that can be useful for decision-making. The work reported in this paper focuses on how to extract and integrate this information in order to construct a Data Warehouse that makes it available. The manual process of extracting and integrating information is expensive and complex. That is the reason why we suggest the development of a tool, based on Wrappers and Mediators, which allows the extraction of information from the Web and integrate it automatically. Wrappers are in charge of information extraction, which is based on page’s structure and a query enriched using a domain ontology. Mediators perform data integration in order to combine information from several sources, solving the conflicts that may appear due to contradictory information, taking into account the trust of the sources. An important characteristic we consider in our proposal is that information contained in the Web changes constantly; therefore a mechanism that supports system evolution becomes essential. For this reason we propose the generation of metadata that keeps the traceability of the process, allowing managing the impact of the source changes on the whole system.