Administering Models: Overview

This document covers how to configure access to models and the modeling integration points between the common model repository and SAS Viya products. The common model repository is associated with the Model Repositories content root folder and can contain one or more repository folders. The Model Repository service provides access to a common model repository for SAS applications, and enables users to perform the following actions:

import models into SAS Model Manager
register models from SAS Model Studio, SAS Visual Analytics, and SAS Data and AI Studio into the common model repository
add a model from the common model repository into a decision flow in SAS Intelligent Decisioning
publish models or decisions to configured publishing destinations
refer to models that are stored in the common model repository from a project within SAS Event Stream Processing Studio

Modeling Integration Points

Here are the integration points for the common model repository:

CAS

CAS stores analytic store files in the ModelStore caslib. Users can also publish models from SAS Model Manager and SAS Model Studio to a CAS publishing destination.

See Data in SAS Cloud Analytic Services: Fundamentals.

Common Model Repository

The common model repository contains multiple repositories that are used to store models that have been registered from SAS Model Manager, SAS Model Studio, SAS Data and AI Studio, and SAS Visual Analytics. The models can then be accessed from other SAS applications, such as SAS Intelligent Decisioning, or published to external publishing destinations.

See Managing Model Repositories in SAS Model Manager: User’s Guide.

SAS Model Studio

A suite of SAS products that enables you to build SAS Viya Machine Learning models as well as SAS Visual Text Analytics models and then to register them into the common model repository. SAS Viya Machine Learning models can also be published from SAS Model Studio to a configured published destination.

See SAS Viya: Machine Learning User’s Guide and SAS Visual Text Analytics: User’s Guide.

SAS Event Stream Processing Studio

A SAS web application that enables you to create, edit, upload, publish, and test event stream processing models using SAS Event Stream Processing Studio Modeler. SAS Event Stream Processing Studio Modeler displays a model as a data flow diagram, which enables you to see and control how windows relate and flow to one another. Projects can reference models that are stored in the common model repository. When a project is deployed, the model is retrieved from the common model repository and written to the ESP server. SAS Micro Analytic Service modules are used to accommodate the imported content that was created in SAS Model Manager. The module is uploaded and then referenced from the Calculate window’s input handler. See SAS Event Stream Processing: Using SAS Event Stream Processing Studio.

SAS Data Governance

A SAS web application that enables you to capture and enrich metadata for files, tables, and other information assets. This metadata is stored in a catalog. You can search the catalog to find assets that you need to meet your business goals. SAS Model Manager can access the asset metadata and analyze the training data referenced on the Model Card tab of a model.

See SAS Data Governance: User’s Guide.

SAS Intelligent Decisioning

SAS web application that enables you to combine analytical models, rule sets, and conditional logic into decisions. You can investigate various scenarios, test and refine the decision logic, and then publish the decisions for use in batch applications and online transactions. After a decision has been published, it is available for use by other applications. See SAS Intelligent Decisioning: User’s Guide.

SAS Model Manager

SAS web application that enables you to store and manage models in a common model repository, as well as to organize them within projects and folders. You can import models that you developed using a SAS application (such as SAS Model Studio, SAS Visual Analytics, and SAS Data and AI Studio) as well as SAS code, open-source programming languages, and PMML. You can also create a new model with the model’s files in a folder or project. Models that are located within a project can be evaluated for champion model selection and monitored for performance. Models within projects can also be published to a configured publishing destination that can be defined for CAS, Hadoop, SAS Micro Analytic Service, and Teradata, as well as Amazon Web Services (AWS), Azure, and Private Docker containers.

See SAS Model Manager: User’s Guide.

SAS Data and AI Studio

SAS development application for running SAS programs, which enables users to use macros to create, update, and delete objects within the common model repository. A user can also use the Register task to import a scoring model from SAS Data and AI Studio into a SAS Model Manager project that is located within the common model repository. A scoring model is an analytic object in a CAS table. Scoring models can be created using several SAS Data and AI Studio tasks such as the Forest task.

See Register SAS Model in SAS Data and AI Studio: Working with Flows.

SAS Visual Analytics

A SAS web application that enables you to explore, discover, and predict using your data. If SAS Visual Statistics is licensed at your site, then you can create, test, and compare models based on the patterns that are discovered during exploration of your data. You can export a model before or after performing model comparison in order to use it with other SAS products in a production environment. If SAS Viya Machine Learning is licensed at your site, then additional models are available. SAS Viya Machine Learning cannot be licensed without SAS Visual Statistics. You can also register SAS Viya Machine Learning and SAS Visual Statistics models from SAS Visual Analytics to the common model repository.

See SAS Visual Analytics: Working with Machine Learning Objects and SAS Visual Analytics: Working with Statistics Objects.

Last updated: September 10, 2026