Forecasting Using Model Studio in SAS Viya

This course provides a hands-on tour of the forecasting functionality in Model Studio, a component of SAS Viya. The course begins by showing how to load the data into memory and visualize the time series data to be modeled. Attribute variables are introduced and implemented in the visualization. The course then covers the essentials of using pipelines for generating forecasts and selecting champion pipelines in a project. It also teaches you how to incorporate large-scale forecasting practices into the forecasting project. These include the creation of data hierarchies, forecast reconciliation, overrides, and best practices associated with forecast model selection.

Learn how to

• Automatically create and fit custom forecast models using structured analytic workflows or pipelines.
• Visualize modeling data using attribute variables.
• Refine forecast models to improve forecast accuracy.
• Apply overrides-generated forecasts.
• Generate forecast data sets for deployment.
• Build and share custom pipelines for large-scale forecasting analyses.

Who should attend

Forecasters and analysts in any industry, including retail, financial services, manufacturing, and pharmaceuticals

Formats available	Standard Duration (duration can vary, see event schedule for details)
Classroom:	2 days
Live Web Classroom:	4 half-day session(s)	System Requirements
e-Learning:	14 hours/180 day license

Although prerequisites are not required, you will get more value from this course if you first attend these two courses:

• Time Series Modeling Essentials.
• Statistics You Need to Know for Machine Learning.

This course addresses SAS Visual Forecasting software.

Introduction and Data Visualization

• Lesson overview.
• Getting started with your forecasting project.
• Visualizing time series using attribute variables.

Pipeline Essentials

• Lesson overview.
• Definition and creation of a time series.
• Fundamental concepts in time series modeling.
• Classes of time series models.
• Model comparison using honest assessment.
• Pipeline templates and pipeline comparison.
• Interactive modeling node.

Hierarchical Forecasting

• Lesson overview.
• Building a hierarchical forecasting model.
• Selecting a champion model.

Post-forecasting Functionality

• Lesson overview.
• Overrides and exporting generated tables.

Creating Better Forecasts and Improving Modeling

• Lesson overview.
• New and existing attributes and filters.
• Data export.
• Score new data and task options.
• Event variables.
• Intermittent data and models.
• Outlier detection.

In-Line Code Access and Overview (Appendix)

• Code overview.
• Demo: Auto-forecasting code overview.
• Demo: Modifying the auto-forecasting code and creating a custom forecast node.
• Demo: Overview and modification of the hierarchical forecasting (pluggable) node.
• Demo: Modifying and saving a hierarchical forecasting (pluggable) node.