Logical Schema Acquisition from Text-Based Sources for Structured and Non-Structured Biomedical Sources Integration

In this paper we present a novel approach to integrate non-structured and structured sources of biomedical information. We part from previous research on database integration conducted in the context of the EC funded INFOGENMED project. In this project we developed the ONTOFUSION system, which provides a robust framework to integrate large sets of structured biomedical sources. Methods and tools provided by ONTOFUSION cannot be used to integrate non-structured sources, since the latter usually lack a logical schema. In this article we introduce a novel method to extract logical schemas from text-based collections of biomedical information. Non-structured sources equipped with a logical schema can be regarded as regular structured sources, and thus can be bridged together using the methods and tools provided by ONTOFUSION. To test the validity of this approach, we carried out an experiment with a set of five cancer databases.