BookmarkSubscribeRSS Feed
cwid05
SAS Employee

We are pleased to announce the most recent release of the SAS Model Implementation Platform (June 2018). In MIP 3.1, we have two large features (Attribution Analysis and Risk Model Editor) along with several additional features that enhance the user experience. We look forward to your feedback and suggestions as we continue to develop the product. We plan to add detailed blog posts (and possibly videos) for many of these features in the coming weeks.

 

  • Attribution Analysis
    • Built-in attribution analysis allows you to select two runs, order the input factors (economic attributes, risk data objects and computed methods) and create an analysis that explains the differences between two runs. We produce a waterfall chart that displays the results and can be saved to a PNG.
  • Risk Model Editor
    • You can now create and edit models with our new UI (instead of creating the model and vars data sets). We now accept SAS7BDAT and CSV formats for coefficients, transition rates and baseline data.
  • Backtesting Metrics
    • We provide five backtesting metrics that can be added to any Backtesting Scenario run. The user can also create their own metrics in the Computed Methods section. We also enable the ability to publish the metrics to Model Risk Management for Modeling System runs.
  • Custom Librefs
    • Each user can establish their own librefs that they can use to select input data from (e.g., portfolio) or write output data to (e.g., Scentrans).
  • Graph Economic Scenarios
    • We allow you to graph your economic data prior to executing your run or after its completion.
  • Cash-flow Legs and Time Bucket Schemes
    • You can now create cash-flow legs in your model group. We also enable the creation and use of time bucket schemes in your run. Note both of these operate on the HP Risk cash
  • Properties Panel and General UI enhancements
    • We have a properties panel that contains the data, settings and selections used to create a run. We have also added information about the size of the cube created by the run.
  • Basecell in Stress
    • You can now use the BASECELL option in stress runs to reduce the footprint of your run. To do so, add Basecell to the PROC HPRISK arguments and remove INSTID from your Cross-classification variables during the creation of your portfolio cube.
  • Multiple Simultaneous Shocks in Sensitivity Analysis
    • We now allow you to create multiple simultaneous shocks in a Sensitivity Analysis run by changing 2 or more shocks to have the same name.
  • Risk Explorer in HTML5
    • The Risk Explorer has been rebuilt in HTML5. It has similar (although not equivalent) functionality.
  • Risk Workgroups
    • We have adopted Risk Workgroups so the file path to the MIP directories has slightly changed.
  • REST API Lua Script
    • We have created wrapped several common REST API calls in Lua functions that can be executed in SAS via PROC LUA. This will help making batch scripts to execute MIP much simpler. We plan to expand this in the future for common functionality.

Please consult the new SAS Model Implementation Platform 3.1 User Guide for detailed descriptions of each of these.

3 REPLIES 3
Alpay
Fluorite | Level 6

Is it backward compatible with v2.4?

cwid05
SAS Employee

Yes we are backward compatible with the SAS Model Implementation Platform 2.4. There is a migration process described in the Administrator's Guide.

Austin9462
SAS Employee

Yes. Any exports from previous versions of the Model Implementation Platform, either individual analysis objects (Model Groups, PostProcess Methods, etc.) or Modeling Systems are backwards compatible and can be imported into later releases of MIP.