A Semantic Enrichment of Data Tables Applied to Food Risk Assessment

摘要

Our method allows one to enrich semantically tables extracted from heterogenous documents found on the Web. The semantic enrichment is completely automatic and is guided by an ontology of the domain. Thus, that processing cannot lead to a perfect and complete enrichment. The SML representation we propose keeps all the possible interpretations, incompletely identified relations and original elements of the context. Contrarily to previous approaches like, we cannot base the search for information on a common structure discovered among a set of homogeneous documents.