Question about the CVmodels generated from regression learner app in machine learning toolbox?

In the regression learner, one can choose the "cross validation" option (i.e., 5 fold) to generate a model and use that model to make predictions using mdl.predictfun(testdata)

However, if done by code, when "cross validation" is specified, the model generated is "RegressionPartitionedEnsemble" type, which cannot be used directly using predict(mdl, testdata).

I'm just wondering, how is the CVmodel generated with the app learner? what is the equivalent way to generate it with code so that one can use to make predictions?

Answer 1

[Amit Doshi]

Hello Xi,

You can generate the code which was used by the app learner to train the model by following the below link:

https://www.mathworks.com/help/stats/export-regression-model-to-predict-new-data.html