One of the most interesting SAS product launches of 2009. I am currently reading SAS Enterprise Miner and I am quite impressed – in fact we have a 1500 processor HPC cluster , besides access to Kraken, the 3 largest HPC in the world. It is interesting to see possible application uses for that. Of course I am currently fiddling with R based parallelized clustering on them.

SAS^® Scoring Accelerator

Citation-

http://www.sas.com/technologies/analytics/datamining/scoring_acceleration/index.html

Quickly and accurately process and score analytic models built in SAS^® Enterprise Miner^TM

What is SAS^® Scoring Accelerator?
SAS Scoring Accelerator translates and registers SAS Enterprise Miner models into database-specific functions to be deployed and then executed for scoring purposes directly within the database. SAS Scoring Accelerator is a separate product that works in conjunction with  SAS Enterprise Miner.

Why is SAS^® Scoring Accelerator important?
SAS Scoring Accelerator automates the movement of the model scoring processes inside the database. Faster deployment of analytic models means more timely results, enabling business users to make important business decisions. Better-performing models help ensure the accuracy of the analytic results you’re using to make critical business decisions.

For whom is SAS^® Scoring Accelerator?
SAS Scoring Accelerator is specifically for organizations that use SAS Enterprise Miner. It is designed for chief scoring officers and IT to score analytic models directly inside the database.

Key Benefits:

 

• Achieve higher model-scoring performance and faster time to results.
• Reduce data movement and latency.
• Improve accuracy and effectiveness of analytic models.
• Reduce labor costs and errors by eliminating model score code rewrite and model revalidation efforts.
• Better manage, provision and govern data.

 

Key Features:

Export Utility:

• Functions as a plug-in to SAS Enterprise Miner that exports the model scoring logic including metadata about the required input and output variables.

Publishing Client:

• Automatically translates and publishes the model into C source code for creating the scoring function inside the database.
• Generates a script of database commands for registering the scoring user-defined function (UDF) inside the database. Scoring UDFs are available to use in any SQL expression wherever database-specific built-in functions are typically used.
• Supports a robust class of SAS Enterprise Miner predictive and descriptive models including the preliminary transformation layer.

SAS Scoring Accelerator interfaces with the following relational databases:

• SAS^® Scoring Accelerator for Teradata
• SAS^® Scoring Accelerator for Netezza