Joshua J. Cogliati

> I dealt with the comments on the doc file, however, I cannot compile the user manual to check whether the changes are displayed correctly. The problem seems to be...

After Input improvements #395 goes in, PostProcessors, Steps, Distributions and DataObjects use the input handling system. Samplers use the input handling system to parse the input, but generally still use...

Tagging @wangcj05 since he is working on ROM improvement, and @Jimmy-INL since he has also requested this ability.

Okay, I see that in framework/SupervisedLearning/PolyExponential.py However, framework/SupervisedLearning/SciKitLearn.py does not seem to have a ```writeXML``` method, so MLPRegressor will not be able to do that at the moment.

Yes, the base method is in SupervisedLearning, but it only does: ``` writeTo.addScalar('ROM',"noInfo",'ROM has no special output options.') ```

For SciKitLearn we could probably do something like (untested psedocode): ```python writeXML(...): for index in range(len(self.ROM)): if "coefs_" in dir(self.ROM[index]): writeTo.addVector("coefs_"+str(index), self.ROM[index].coefs_) if "intercepts_" in dir(self.ROM[index]): writeTo.addVector("intercepts_"+str(index), self.ROM[index].intercepts_) ``` It...

Yes, possibly after #1607 it might be as simple as each class that inherits from ScikitLearnBase has a list of scalars to export and vectors to export.

LinearRegression with actual tested code: https://github.com/idaholab/raven/commit/ab4e3e76c5f567aeded53e55e6442b0f06774be4 ``` diff --git a/ravenframework/SupervisedLearning/ScikitLearn/LinearModel/LinearRegression.py b/ravenframework/SupervisedLearning/ScikitLearn/LinearModel/LinearRegression.py index a1e527ab1..8526baeb3 100644 --- a/ravenframework/SupervisedLearning/ScikitLearn/LinearModel/LinearRegression.py +++ b/ravenframework/SupervisedLearning/ScikitLearn/LinearModel/LinearRegression.py @@ -86,3 +86,17 @@ class LinearRegression(ScikitLearnBase): # notFound must be empty assert(not...

Maybe it could be done this way: ```python #in LinearRegression.py self._vectorWriteList = ["coef_"] self._scalarWriteList = ["intercept_"] #in ScikitLearnBase.py def writeXML(self, writeTo, targets=None, skip=None): """ Allows the SVE to put whatever...

Note that this was partially fixed for covariance type full, but not fixed for type diag. See XXX comment in framework/unSupervisedLearning.py ```python #if covariance type == full, the shape is...