SAS® Forecast Server
What does SAS Forecast Server do? SAS Forecast Server generates large quantities of high-quality forecasts quickly and automatically, allowing organizations to plan more effectively for the future.
Why is SAS Forecast Server important? The unsurpassed scalability of SAS Forecast Server enables all levels of your business to operate more efficiently by quickly and automatically producing millions of high-quality forecasts you can trust.
For whom is SAS Forecast Server designed? It is designed for organizations in any industry that need large-scale forecasting and/or require automation because of the large number of forecasts or a lack of highly-analytic forecasters.
Key Benefits
- Provides forecasts in a quick and timely manner through a user-friendly graphical interface. SAS Forecast Server automatically produces high-quality forecasts with the ability to modify models interactively without programming. This makes large forecasting processes manageable and allows analysts to focus their time on the most important forecasts.
- Provides forecasts that reflect the realities of the business, improving your ability to plan future events with confidence. Only SAS Forecast Server automatically selects the business drivers, holidays or events that aid in the forecasting process from any number of variables supplied to the system in the modeling process. Forecasts better reflect the business and require less overriding and fewer manual interventions. You can trust the accuracy of your forecasts.
- Improves forecasting performance across all products, inventory levels and stock-outs. SAS has a complete array of advanced forecasting methods and can statistically estimate the impact of sales and marketing events based on sound business logic. Graphical displays of the impacts provide a greater understanding of the effects of holidays, marketing events, sales promotions and unexpected events, such as weather, improving the ability to forecast and plan future sales promotions and marketing events.
Key Features
- New Project wizard. The wizard enables novice forecasters to set up the automatic forecasting process quickly and easily. It guides users through data selection, assigning roles to variables in the data set, setting up a forecasting hierarchy and selecting important automatic forecasting criteria.
- Automatic forecasting. SAS Forecast Server intelligently determines which forecasting models best fit the historical data. Model parameters are automatically optimized to provide the most appropriate model, resulting in more responsive and accurate forecasts. Holdout samples can specified so that forecasting models are selected not only by how well they fit the past data, but by how well they are likely to predict the future. Users can choose the level of automation of the forecasting process.
- Hierarchical forecast reconciliation and disaggregation. Top-down, middle-out and bottom-up forecast reconciliation is provided to support the hierarchical nature of many forecasting processes. Statistical forecasts at the lowest level of the hierarchy can be aggregated to create forecasts at higher levels. Similarly, statistical forecasts at higher levels of the hierarchy can be allocated to lower levels. Forecasters can specify a forecasting hierarchy and the default options within the New Project wizard.
- Exception rule settings. Automatic forecasting processes are not always perfect. SAS Forecast Server lets users set up business rules for flagging potentially problematic forecasts. Upon completion of the automatic forecasting process, forecasters can quickly identify forecasts that violate a defined rule so they can focus attention where it is most needed.
- Events management console. Events such as sales promotions, unusual weather, etc., can greatly affect forecasts. An events management console allows users to create event definitions, assign events to selected series in the project and delete events. Users can specify event duration, shape and recurrence options. Predefined common events and holidays are available, making model development and deployment fast and easy.
- Code generation for batch processing. SAS Forecast Server generates SAS code through the interactive graphical interface. Users can export the code to edit the project in a program editor, schedule and run projects in a batch mode or create SAS Stored Processes.
- User override facility. Forecasters can override the statistical forecast to incorporate judgment or outside information into the forecasting process. Forecasters who do not wish to refine the forecasting models may find that overriding forecasts is the quickest way to deal with problematic forecasts.
- Client/server architecture. SAS Forecast Server can reside on a single machine or on a server for multiple users to access. The client/server environment makes SAS Forecast Server suitable for large-scale forecasting problems. SAS Forecast Studio is a Java-based GUI, which can be set up to connect to SAS Forecast Server either on a server or on the same standalone machine.
- Optimized model parameters. Mathematically optimized model parameters are provided so users don't have to guess and manually enter model parameters or perform a cumbersome grid search for reasonable estimates. Optimized parameters provide better fitting models and forecasts that are more accurate and responsive to changes in the data.
- Automatic regressor/event selection and model specification. Regressors and events that improve the forecasting model are automatically selected from any number of regressors supplied to the system, and SAS Forecast Server automatically determines how they are specified in the model. The system examines the contemporaneous relationships of the regressors and events to the items being forecast and determines whether lagged and/or dynamic relationships are present. It automatically computes variable transformations, lags and transfer function definitions.
- Automatic outlier detection. SAS Forecast Server examines the history of each item being forecast and automatically identifies outliers and shifts in the data. Subsequent forecasts adjust for these outliers and shifts appropriately.
For a complete list of key benefits and features, refer to the SAS Forecast Server Fact Sheet. 
E-mail this page 
Print this Page 
Digg this 
Save to del.icio.us 
RSS feed 