Microsoft FAST Search Server for SharePoint, SA, OLP-NL AEF-00078 Manuel D’Utilisation
Codes de produits
AEF-00078
16
www.microsoft.com/sharepoint
improve the precision of the property extraction by editing the include lists and exclude
lists for each property extractor. Excluded items are removed immediately, while
included items take effect the next time the content is indexed. You can also create
custom property extractors based on your organization's specific content using
Windows PowerShell and SharePoint administration. Dictionary or Taxonomy based
extractors, also called verbatim extractors, will allow you to extract managed properties
based on a fixed list of known terms. Developers will also be able to create more
dynamic extractors based on the FAST matcher framework and will be able to extend
the document pipeline with specialized classifiers, entity extractors, or other processing
can be used to support specialized scenarios.
lists for each property extractor. Excluded items are removed immediately, while
included items take effect the next time the content is indexed. You can also create
custom property extractors based on your organization's specific content using
Windows PowerShell and SharePoint administration. Dictionary or Taxonomy based
extractors, also called verbatim extractors, will allow you to extract managed properties
based on a fixed list of known terms. Developers will also be able to create more
dynamic extractors based on the FAST matcher framework and will be able to extend
the document pipeline with specialized classifiers, entity extractors, or other processing
can be used to support specialized scenarios.
Rank Profiles
The index schema in FAST Search Server 2010 for SharePoint includes rank profiles,
which control how relevancy ranking is calculated for each item in search results.
which control how relevancy ranking is calculated for each item in search results.
A rank profile defines how relevancy calculations are performed when you search a full-
text index. A rank profile consists of several components which are weighted when
calculating an item's relevance. You can adjust the weights of a profile‟s components to
improve search result relevance. Rank profile components include the following:
text index. A rank profile consists of several components which are weighted when
calculating an item's relevance. You can adjust the weights of a profile‟s components to
improve search result relevance. Rank profile components include the following:
Freshness. This component manages how the age of an item affects rank.
Proximity. This component manages how the distance between query terms
affects rank.
affects rank.
Authority. This component manages how links between Web documents affect
rank.
rank.
Query authority. This component manages how user selections in previous query
results affect rank.
results affect rank.
Context. This component manages how different managed properties within the
associated full-text indexes contribute to the rank.
associated full-text indexes contribute to the rank.
Managed properties directly impacting the rank. You can specify that the value of
a numeric managed property is added to the rank, or you can specify that certain
values of a managed property impact the rank. In the latter case you can, for
example, define that documents of a given type (as defined by a specific
managed property) will get a relevancy boost in the results.
a numeric managed property is added to the rank, or you can specify that certain
values of a managed property impact the rank. In the latter case you can, for
example, define that documents of a given type (as defined by a specific
managed property) will get a relevancy boost in the results.