In the July 2015 release
of SAS 9.4, the SAS Scoring Accelerator for Hadoop, SAP HANA, and
Teradata support complex model scoring, using analytic stores. With
complex models, the traditional descriptive code is no longer adequate
to express such models due to the huge amount and variety of information
that is needed. Analytic stores present themselves as the alternative
to the old style of scoring with code generated by the underlying
engine.
Three files are required
for analytic store scoring by the SAS Scoring Accelerator:
-
the analytic store file
An analytic store is
a binary file that contains information about the state of an analytic
object. It stores information that enables you to load and restore
the state of the analytic object and set it in a score-ready mode.
The analytic store is transportable. That is, it can be produced on
one host and consumed on others without the need of the traditional
SAS export or import.
By default, the filename
is score.sasast. However, any filename can be specified.
-
the score code
The score code is a
DS2 program.
By default, the filename
is score.sas. However, any filename can be specified.
-
an XML file that contains the scoring
variables and other properties that are used and created by the scoring
code
By default, the filename
is score.xml. However, any filename can be specified.
For information about
how to use the SAS Scoring Accelerator to perform analytic store scoring,
see the SAS Scoring Accelerator section for your data source.