What TREX index types activate witch services?

[Former Member]

Hi,

I see from Karstens reply () to an earlier discussion that a pure "search" index disables classification, similarity searching, similar terms and "etc".

So I guess that the "classification" type index activates the services:

- classification

- similarity searching

- similar terms

What services are "etc", and what services does the "search" index activate? Is there a document explaining this, and what the different services does?

Basically: What services does the "search" index activate/utilize and what services does the "classification" index activate/utilize, and a description of the services.

Best regards,

Bjorn

[KarstenH]

Hi Bjorn,

the "etc." points to anything that could also be based on the text-mining engine of TREX. Anything that works with text similarities and/or statistic associations of terms.

In the TREX INdex Server there are three sub-engines:

- The Indexing Engine that builds up search indices and performs any pure searches over document content (including fuzzy and lingusitic)

- The Text-Mining Engine that builds up a multidimensional vector space from all indexed terms and documents. It determines document similarities, extracts related terms and does example-based classification. It is disabled (including the query-based classification, although that is not based on the vector engine) when chosing a pure search index.

- The Attribute Engine that indexes and searches over KM properties and is also heavily used in other (not KM), table-data-focused scenarios that use TREX.

You can find a presentation that may illustrate this here post under:

http://service.sap.com/~sapidb/011000358700004850302004E/TREX_for_SAPTechEd_2004.ppt

<b>NOTE:</b>

Not all TREX technical abilities that are described here are fully available in KM.

Regards, Karsten