SCA WorkBench 5.2

A companion product to the PC SCA Statistical System that provides forecasting and data analysis links to popular spreadsheets!

SCA WorkBench provides an intuitive and easy-to-use graphical user interface (GUI) for applications in time series analysis and forecasting. It also provides awesome features to benchmark model performance using rolling forecasting techniques, handle large amounts of data through links to popular spreadsheets, automate analysis and forecasting procedures through task and project grouping, and much more.

SCA WorkBench enhances your ability to automate various analysis tasks through SCA macro procedures. It provides a variety of utilities to organize macro procedures by task and group multiple tasks into a project.

SCA WorkBench provides an internal mechanism to link an analysis task to a data source. This is a tremendous benefit if you manage your data in spreadsheets and require a convenient way to access the data in the PC SCA System. SCA WorkBench provides a seamless link into Microsoft Excel, Lotus 1-2-3, Corel Quattro Pro, dBase, Clipper, FoxPro, as well as tab-delimited text files. Other data formats may be included in future updates.

SCA WorkBench allows you to run an unlimited number of tasks in a session. If you specify a data link with a spreadsheet, SCA WorkBench gives you the option to include any new data in your spreadsheet before it runs the analysis task. This allows applications involving large numbers of data series to be handled easily within this general framework. Data may be maintained and managed directly in your spreadsheet avoiding the hassle of updating data in more than one data source.

SCA WorkBench runs on Windows 95/98, NT, and Windows 2000 operating systems. SCA WorkBench uses high speed external DLLs and the latest Object Linking and Embedding (OLE) technology to control access to various spreadsheet, word processing, and utility programs from a central work area.

SCA WorkBench may be customized to handle a variety of decision support applications. If you have a need to automate a decision support application, contact SCA Consulting and Support Services for more information.

SCA WorkBench provides an intuitive and easy-to-use graphical user interface (GUI) to build ARIMA, transfer function and intervention models using an automatic time series modeling or user-defined modeling environment.

SCA WorkBench provides comprehensive features to benchmark time series models and evaluate forecasting performance using rolling forecasting. Also offers advanced capabilities to evaluate forecasting performance in the presence of outliers. Various performance criteria available in rolling forecasting include standard and robust calculations of root mean square error (RMSE); mean absolute percentage error (MAPE); mean absolute deviation (MAD); and and cumulative sum of errors (CUSUM). These results can then be exported to an Excel spreadsheet or standard text file.

SCA WorkBench enhances connectivity between the PC SCA System and Microsoft Excel. SCA WorkBench allows you to write SCA results into a an Excel spreadsheet.

SCA WorkBench now includes an automatic task and macro building capability. Using this feature, you can specify a master SCA macro file that is to be applied to all variables (or selected variables) in a data source. All tasks are then automatically organized in a project file that can then be executed in a batch manner. This capability also supports macro replication that allows users to define do loops in a master macro. As the tasks and macro files are being generated, the do loops defined in the master macro procedure automatically replicate (or expand) SCA commands.

SCA WorkBench provides links to programs that are developed in the SCAB34S applet development environment. The SCAB34S product offers a comprehensive mathematical and statistical programming environment. The SCAB34S development toolkit comes with hundreds of pre-programmed subroutines, functions, and programs addressing applications using general nonlinear model estimation, constrained nonlinear programming, matrix programming, GARCH-M model estimation, and much more. Please contact SCA if you are interested in purchasing the SCAB34S applet development environment. No compiler is required.

User-specified programs and other off-the-shelf programs can now be added dynamically to the Program menu of SCA WorkBench. Once a program is added to the Program menu, it can be executed directly from SCA WorkBench.

The help system for SCA Statistical System syntax has been improved in SCA WorkBench. All SCA commands are categorized by function allowing users to find a specific command easily. Users can also search, print, edit, reorganize, add and delete SCA help files.

Help for SCA applet programs (e.g., XLSREAD, XLSWRITE, and SCAB34S) are also accessible from SCA WorkBench. The help facilities are designed in a manner such that users can build help files for other programs accessed by SCA WorkBench. This includes programs that were developed using the SCAB34S applet development environment, SCA applet development toolkit, or any off-the-shelf software product. As an added convenience, the user can automatically load and execute sample programs that are associated with a help file. Because the help system is extensible, users can add additional sample programs and share these programs among other users.

Univariate Time Series Model Extraction Engine: With the model extraction engine in SCA WorkBench, users can parse multiple SCA output files and extract information related to ARIMA, transfer function, and intervention models. The model information is then automatically collated and tabulated in a tab-delimited summary table. The summary table may then be imported into any spreadsheet or database application. Among the information that may be parsed include source filename; model instance; model name; dependent variable name; differencing operators; residual standard error; AIC and SIC; parameter estimates, standard errors, and t-values; and more.

SCA WorkBench provides a convenient graphical user interface (GUI) for users interested in estimating conditional heteroscedastic (ARCH/GARCH) models. Through the ARCH/GARCH modeling environment, users can select various types of models under a simple GUI and automatically submit those models for estimation in the SCA System or SCAB34S NONLIN module. Current models supported include ARCH, GARCH, GARCH-M, Exponential GARCH, Integrated GARCH, estimation using student-t distribution (FATTAIL) and GJR GARCH.