Gordana Stojsic, Igor Ljubi,
FER, University of Zagreb
Using
metadata in Web pages has been recognized as one of the ways
to make Web search more efficient. In the area of metadata
standardization, the Dublin Core metadata standard has been
developed for describing a wide range of networked resources.
The Croatian Academic and Research Network recommends the
use of Dublin Core metadata in order to improve the searching
of the Croatian Web space.
Metadata can be added to a HTML document manually, by editing
the source, but this approach, aside from being tedious and
time-consuming, may also introduce unintentional errors in
typing, HTML syntax, etc. In order to make adding metadata
easier, a tool for editing Dublin Core metadata was created.
This tool allows introducing new metadata, as well as viewing
and modifying existing metadata in an HTML file. Finally,
the tool also automatically extracts potential keywords from
the text, thus making the selection of keywords easier for
the Web administrator. The keyword extracting algorithm is
based on the frequency of words in the text as well as their
position and formatting emphasis in the text.
This tool was tested on the content of the WWW.HR information
on Croatia covering the topics history, economy, culture,
tourism, nature etc.