Integration of information from multiple sources of textual data.

The chapter presents two ongoing projects towards an intelligent integration of information. They adopt a structural and semantic approach TSIMMIS (The Stanford IBM Manager of Multiple Information Sources) and MOMIS (Mediator environment for Multiple Information Sources) respectively. Both projects focus on mediator based information systems. The chapter describes the architecture of a wrapper and how to generate a mediator agent in TSIMMIS. Wrapper agents in TSIMMIS extract informations from a textual source and convert local data into a common data model; the mediator is an integration and refinement tool of data provided by the wrapper agents. In the second project MOMIS a conceptual schema for each source is provided adopting a common standard model and language The MOMIS approach uses a description logic or concept language for knowledge representation to obtain a semiautomatic generation of a common thesaurus. Clustering techniques are used to build the unified schema, i.e. the unified view of the data to be used for query processing in distributed heterogeneous and autonomous databases by a mediator.