Uncertainty Plug-in
The Uncertainty plug-in provides support for performing uncertainty analysis with SNAP models. The plug-in provides a parametric stream type tailored for uncertainty analysis. This stream type defines parametric tasks based on a set number of random samplings. The random samples are defined in the stream type properties and mapped to model variables: parametric tasks are generated for these each set of random samples. The sequence of analysis code steps that utilize the random variates is entirely defined by the user. Once this sequence of steps is complete, the results are fed into an extraction step, which utilizes AptPlot to compute figures-of-merit. Finally, the figures-of-merit are fed into an uncertainty quantification step used to assess the effect of random variation on the model.
The Uncertainty Plug-in User's Manual provides instructions for working with Uncertainty Job Streams in SNAP. This Manual is also available as a searchable help set in SNAP.
Change Log
Version 1.8.5 - Released 4/17/2025:
- An issue was resolved which prevented executing the APTPLOT extract data phase in the uncertainty analysis on certain clustering software systems.
- An issue was resolved that would prevent submitting a report job step when copied to a new model. Issue 107
- The job stream validation behavior for the uncertainty report job step has been updated to validate the plot definitions.
- When a legacy uncertainty report job step is opened that contains the old style plot format, the plots will be converted to the new format.