- In this paper we present a suite of tools to automatically
acquire and browse conceptual schemas from large
collections of HTML-based biomedical documents. This
suite is composed of two tools: the schema acquisition
tool (SAT) and the zoomable browser (ZB). The SAT is
the implementation of a novel four-phased method to
extract conceptual schemas from non-structured sources.
First, all documents in the collection are analyzed to
extract relevant concepts. Second, the vocabulary
discovered during the first phase is organized into a
hierarchical structure. Third, the schema is enriched with
non-hierarchical ad-hoc relationships. The last phase is an
optional refinement activity that must be conducted by
experts in the domain covered by the collection. The
extracted schemas can be navigated using the ZB. We
have used these tools for different purposes in the EC
funded biomedical research project Advancing Clinico-
Genomic Trials on Cancer (ACGT), obtaining promising
results.
(from crpit.com)
(local if available)
