What's New in SAS Model Manager 3.1

Overview

SAS Model Manager 3.1 has the following new features and enhancements:
  • ability to retrain models
  • ability to customize dashboard reports
  • enhanced support for modifying project definitions
  • ability to import R models
  • ability to modify and upload templates and SAS files
  • ability to manage the progress of a project or version using SAS Workflow
  • enhanced support for PMML models
The first maintenance release of SAS Model Manager 3.1 adds the ability to publish scoring functions to a Greenplum database.

Ability to Retrain Models

SAS Model Manager now supports the retraining of models. Multiple attributes can be specified when you define a model retrain task. You can also register the models and create model comparison reports when executing a model retain task for a project.For more information, see Overview of Retraining Models.

Ability to Customize Dashboard Reports

The SAS Model Manager Dashboard reports are produced from the same data sets that are used to create the performance monitoring reports. The data sets are created by running performance tasks. For each project a user can define dashboard report indicators that are used to create the dashboard reports. Users can now specify additional report options and exclude project types when generating the dashboard reports. The dashboard reports are not displayed through the SAS Model Manager user interface, but instead are viewed in an HTML browser. For more information, see Overview of Project Dashboard Reports.

Enhanced Support for Modifying Project Definitions

SAS Model Manager administrators can now modify the project definition that contains project properties, project input variables, and project output variables, after a project has been created. This change enables users to modify the project properties, select new project input and output variables, or modify existing project input and output variables. A SAS Model Administrator can also lock the project metadata so that it cannot be modified. For more information, see Lock or Unlock Project Metadata and Modify Project Definition.

Ability to Import R Models

SAS Model Manager now supports importing R models. R is a freely available language and environment for statistical computing and graphics. Using the open architecture of SAS Model Manager, you can register and import R models. For more information, see Import SAS Code Models and R Models Using Local Files.

Ability to Manage Templates and SAS Code Files

Using the SAS Model Manager Template Editor, you can create, edit, or delete life cycle templates, model templates, and user report templates or user report SAS code files. You can then upload new or modified templates, or SAS code files to the SAS Content Server. For more information, see Creating Life Cycle Templates, User-Defined Model Templates, and User-Defined Reports.

Ability to Manage the Progress of a Project or Version Using SAS Workflow

The Workflow Console for SAS Model Manager can be used to track the progress of a modeling project or version. Process definitions are created using SAS Workflow Studio and are activated with the SAS Workflow Engine. A SAS Model Manager administrator can then use the Workflow Console to create instances of the process definitions to be used with SAS Model Manager. For more information, see Overview of Workflow Console.

Enhanced Support for PMML Models

SAS Model Manager now supports setting a PMML model as the champion model, publishing a PMML model to a SAS channel, and exporting a PMML model to the SAS Metadata Repository.

Ability to Publish Scoring Functions to a Greenplum Database

In the first maintenance release of SAS Model Manager 3.1, you can publish scoring functions to a Greenplum database. SAS Model Manager uses the scoring publishing macros that are included in the SAS/ACCESS Interface to Greenplum to publish the scoring information for a model to the database. For more information, see Publish Scoring Functions.