Semantic technology is based on the use of ontologies, which are nothing more than sets of definitions or concepts and relationships between them (a structured collection of information) that we use to describe things (domains).
Basically, what we do is put autotagging marks (labels) within documents that correspond to categories included in an ontology to improve access and retrieval of documentation (search).
As I have said before, the power is in identifying which terms are relevant and defining whether it is worth to convert them into tags. With very global ontologies or generic terms, the result might not be entirely relevant to our context. As we saw in the resume example, a proposed tag was "Europass", but maybe, what we really wanted to know was which candidates had certain knowledge about computer technology, so "Europass" told us nothing. It is necessary, therefore, that we describe a domain or a narrow spectrum.
This is precisely one of the improvements that have been made in the latest version of Autotagging. Today it is possible to ask Athento to label with the categories defined in our ontology, we only need to load the new ontology in the manager.
On the other hand, it has been added the ability to send label proposals to a "blacklist", which indicates to Athento not to use those terms as labels. It is also possible to indicate "synonyms" which should be replaced by a term when it is found, so that we ensure we keep under the same category terms that mean the same thing.
Lastly, but not less important, another improvement is the possibility of constructing document types using labels. A document is recognized under certain type when it has been tagged with a particular set of labels. This new categorization of documents allows us a new formula for virtual navigation of documents. Althoug those documents are stored in different locations, we see them grouped under the same type of document.
Comparing Document Capture Solutions (Athento, Kofax, Ephesoft etc.)
Document Management Success Case at BBVA, managing 7 million records.
Comparing ECM Systems (including Alfresco, OpenText, Documentum, Filenet, Sharepoint or Nuxeo).