Modeling Tab

In this section:

You can use the Modeling tab to access WebFOCUS RStat, a statistical modeling workbench that is embedded in App Studio. You can perform common statistical and data mining tasks, and develop models that can be deployed as scoring applications on every platform. RStat enables data miners and Business Intelligence developers to collaborate with the same tools used to access, manipulate, or transform data, develop predictive models, and create and deploy scoring applications along with associated reports to any worker within their organization.

You can pass any data generated from the Report canvas directly into RStat. After you pass the data to RStat, you can develop your data, join various data sources, build virtual fields, and create data sets to pass to RStat for modeling. The Launch button extracts the data defined within your report and loads it directly into RStat. Within each report, you can customize the name of the hold file that you create, and the location at which it is stored. If you do not customize the name and location of the hold file, a default name and location are assigned.

Note: The Modeling tab is only accessible when RStat is installed.


Top of page

x
Modeling Group

In this section:

How to:

You can run RStat and edit the RStat configuration model options. You can also conduct sampling activities. The commands are:

Launch

Extracts the data defined in your report and loads it directly into RStat. RStat opens in a new window.

Note: The file name that loads in RStat is that which you define on the RStat Model Configuration dialog box. The default file name is HOLD (HOLD.CSV).

Parameters

When you click the Parameters button, the RStat Model Configuration dialog box displays. In this dialog box, you can name the hold file that will be used to pass the model data from the Report canvas into RStat (the default file name is HOLD). You can also define the directory in which the hold file should be placed. You can customize the default hold path for your procedures (.fex files). Your model configuration parameters are saved when you exit the RStat Model Configuration dialog box.

Sampling

Opens the Sampling dialog box. You can generate statistically representative data extracts for data examination and modeling purposes. You can select Random, Numeric Stratified, or Categorical Stratified sampling.

You can also disable sampling. Each option contains a set of criteria that you can edit for the sampling.

Note: In order for sampling to work, you must select a data field prior to clicking Sampling. If you do not select a data field, the message No field is available for sampling will be the only item that appears in the dialog box.



x
Procedure: How to Launch RStat
  1. Create a report in App Studio.
  2. Add fields to the report.
  3. On the Modeling tab, in the Modeling group, click the RStat icon.

    RStat launches in a separate window.

  4. Optionally, customize the variables that are imported into RStat.

    Note: The fields that you define on your report canvas are imported into RStat in the order specified in your report.

    For more information on using RStat, refer to the Using WebFOCUS RStat for Predictive Analytics manual.



x
Procedure: How to Set Parameters Using the RStat Model Configuration

You use the Parameters option to specify a name for the data file, as well as an output directory.

  1. On the Modeling tab, in the Modeling group, click Parameters.

    The RStat Model Configuration dialog box displays.

  2. In the Hold File Name field, enter a name for the file or optionally browse to select an existing file.

    Note: The default Hold file name is HOLD and displays in the file name field on the Data tab in RStat. If you change the name of the file on the RStat Model Configuration dialog box, this revised file name is reflected in RStat.

  3. In the Hold File Path field, enter or select a path in which to save the file.

    Note: You can restore the original drive and directory by clicking Default.

  4. Click OK.

    When you launch RStat, this file name is the default Hold file used upon entry into the application.



x
Sampling Functionality

The sampling functionality is accessible through the Modeling tab in App Studio. Sampling offers three sampling techniques: Random, Numeric Stratified, and Categorical Stratified.

Accessing sampling criteria within the Modeling tab requires that at least one field is defined within the report extract. If no fields have been defined, you will receive a warning message informing you that no fields are available for sampling and the Sampling tab will be present, but no actions will be available. For Random Sampling, which is not tied to a specific field value, any field will open the option. For Stratified sampling, the stratum field (the field on which the strata will be defined) must be included in the report.



x
Basic Sampling Concepts and Terminology

To produce a valid output sample, each sampling definition requires information about the file to be sampled. In each sampling routine, you can use the sample calculator to calculate the values for each of the following or enter your own values:


WebFOCUS