|
Support: auto-classification
Eidetica employs auto-classification for its clients in
two fashions:
-
To support advanced features in search pages, such as
more-like-this and e-mail alerting.
The result of the auto-classification is not sent back
to you but used internally by Eidetica.
-
For FTP-to-FTP enrichment of your XML data streams.
In this scenario, the auto-classifier returns relevance
assessments about the subject classes it assigns.
The auto-classifier links a vocabulary to each
of the subject classes you provide in the training data.
This vocabulary typically consists of tens of thousands
of single or multi-word terms.
Adding or removing subject classes is done by making mutations
of your training set. Eidetica does not keep a separate
copy of your "thesaurus" unless your Service Level Agreement
specifically contains support for the "direct thesaurus" model.
|