RESOURCES / ANALYTICS TOOLS

SAS TO PMML CONVERTERTM

LatentView's SAS-TO-PMMLTM is a FREE desktop application that generates PMML representations for your SAS models.

What is PMML?

PMML (Predictive Model Markup Language), an XML-based language developed by the Data Mining Group (DMG), provides a standard way to represent data mining models so that these can be shared easily across different business intelligence and statistical packages, significantly enhancing the ease of implementation.

PMML allows users to develop models within one vendor's application, and use other vendors' applications to visualize, analyze, evaluate or otherwise use the models.

Why Convert SAS Models to PMML?

Representing SAS models in PMML form gives developers on the SAS® platform the flexibility to deploy models using a whole range of applications that can "consume" PMML files, thus broadening the choice of available deployment platforms, and helping users leverage the capabilities of their existing BI platform investments.

Thus, you could for instance leverage the analytical capabilities of SAS to develop predictive models, convert them to PMML files using LatentView’s SAS-TO-PMMLTM application, load the PMML files to DW/BI applications like MicroStrategy / IBM Infosphere Warehouse, which can then be configured to produce lead lists, reports, dashboards, etc basis these models, for use in various business processes (such as marketing campaigns, collections process, underwriting, etc).

The application is targeted at users of SAS® v8/v9, as these versions of SAS do not have inbuilt functionality to export models to PMML.

SAS-TO-PMML ConverterTM Features

Converts SAS model datasets (EST-type datasets containing parameter estimates) to PMML, as per PMML v3.2 specification
Can handle OLS regression models as well as binary/ordinal logistic regression models
Conversion process is fully automated and UI-driven - requires no knowledge of PMML
OLS Regression Models:
Can convert EST datasets containing multiple models and/or dependent variables in a single run
Logistic Regression Models:
Can handle different types of predictor variables (categorical and continuous predictors, as well as interaction terms)
Can handle various parameterization approaches for categorical variables
Can handle different types of link functions

System Requirements

SAS v8/v9 (only the BASE SAS® software is required)
JRE version 1.6

  Download SAS-2-pmml™