WebFOCUS RStat is a data mining and predictive modeling
workbench embedded in WebFOCUS Developer Studio. RStat supports
common statistical and data mining tasks and the development of
predictive models that can be deployed as scoring applications on
any WebFOCUS platform. RStat enables data miners and Business Intelligence
developers to collaborate with the same tools to access, manipulate,
and transform data, develop predictive models, and incorporate this
scoring functionality into operational Business Intelligence applications.
Data models include decision trees, neural networks, linear and logistic
regressions, random forests, support vector machine, boosting, association
rules, K-means, and hierarchical clustering.
The latest release of RStat 1.3.1 includes the following:
- Model Export C files compliant for Teradata UDF (User Defined
Function) for in-database processing.
- The new Teradata UDF
export function in RStat generates a C program that is Teradata
UDF compliant for in-database processing. When scoring large amounts
of data, running the predictive model using in-database processing
results in significant performance gains.
- After selecting Teradata UDF as an export option, a C file and
a CREATE FUNCTION template are created for the purpose of creating
the Teradata UDF, in a directory of your choice. Using the template
and following the UDF creation process, the UDF is created in Teradata
that is equivalent to the RStat deployed model in WebFOCUS. The
defined function in Teradata can then be used from native Teradata
clients, BI, and ETL.
- Ability to load R scripts to the RStat GUI.
- The Rscript
button in RStat GUI can load an R script with all models that R
supports. If a new package is required, the new package is loaded
accordingly for specific model and code executions. Files or plots
are outputted in the R console or a new window, and then saved to
the directory that is specified in the R script.
- For models that are already implemented in RStat, C routines
can be exported. These models include Regression, Decision Trees,
Clusters, Survival Model, and Market Basket.