The Internet has become a favoured medium for the presentation and exchange
of environmental and chemical data. To search for relevant information, th
e user either has to know the direct address of the Internet site, or has t
o use search engines and meta information repositories. In the latter case,
the desired resource is described by a number of keywords, or descriptors.
However, if too few descriptors are given, the answer set is immensely lar
ge. If too many or too specific descriptors are given, valuable information
might be sorted out, because it lacks a particular descriptor. The Intelli
gent Cluster Index (ICIx) technology can remedy this situation. It generate
s a clustering of documents by their content characteristics. Applied in th
e described scenario this results in a grouping of Internet resources with
comparable content. ICIx offers a similarity search facility based on the c
lustering. It allows the search for an arbitrary combination of descriptors
. If an exact match is required, the result contains only documents matchin
g all descriptors. In the similarity search, documents with comparable cont
ent - identified by the similarity clustering - can be included in the resu
lt set, even if they do not match all descriptors. Thus ICIx offers a wider
range of relevant information in the answer than standard full text search
provides.