Forecasting Using SAS Software: A Programming Approach

This course teaches analysts how to use SAS/ETS software to diagnose systematic variation in data collected over time, create forecast models to capture the systematic variation, evaluate a given forecast model for goodness of fit and accuracy, and forecast future values using the model. Topics include Box-Jenkins ARIMA models, dynamic regression models, and exponential smoothing models.

Learn how to

• Build simple forecast models.
• Build advanced forecast models for autocorrelated time series and for time series with trend and seasonality.
• Build forecast models that contain explanatory variables.
• Build models to assess the impact of events such as public policy changes (for example, DUI laws), sales and marketing promotions, and natural or man-made disasters.

Who should attend

Scientists, engineers, and business analysts who have the responsibility of forecasting or evaluating policies and practices for their organizations

Formats available	Duration
Live Web Classroom:	6 half-day session(s)

Before attending this course, you should have:

• Experience using SAS to enter or transfer data and to perform elementary analyses, such as computing row and column totals and averages, and producing charts and plots. You can gain this experience by completing the SAS Programming 1: Essentials course.
• Experience in data analysis and statistical modeling. You can gain the prerequisite knowledge by completing the Statistics 2: ANOVA and Regression course.
• Experience with stationary ARMA models and elementary forecast models like time trend models and exponential smoothing models for forecasting. You can gain this experience by completing the Time Series Modeling Essentials course.

Knowledge of SAS Macro language programming is useful but not required.

This course addresses SAS/ETS software.

Introduction to Forecasting

• Time series and forecasting.
• Introduction to forecasting with SAS software.
• Evaluating forecasts.

Stationary Time Series Models

• Introduction to stationary time series.
• Automatic model selection techniques for stationary time series.
• Estimation and forecasting for stationary time series.

Trend Models

• Introduction to nonstationary time series.
• Modeling trend.
• Alternatives to PROC ARIMA for modeling trend.

Seasonal Models

• Seasonal ARIMA models.
• Alternatives to PROC ARIMA for fitting seasonal models.
• Forecasting the airline passengers data.

Models with Explanatory Variables

• Ordinary regression models.
• Event models.
• Time series regression models.