MATLAB Answers

How do I find the parameters in discriminant learner of ensemble classification?

1 view (last 30 days)
Bohan Xing
Bohan Xing on 5 Nov 2018
Edited: Don Mathis on 7 Nov 2018
I am using the classification app to decode some information. I extracted code and found the following parameters gave me a good accuracy.
classificationEnsemble = fitcensemble(...
predictors, ...
response, ...
'Method', 'Subspace', ...
'NumLearningCycles', 30, ...
'Learners', 'discriminant', ...
'NPredToSample', subspaceDimension, ...
'ClassNames', [1; 2; 3; 4]);
I know that there are many different discriminant type such as linear or quadratic. And there are hyperparameters such as delta and gamma. (https://www.mathworks.com/help/stats/templatediscriminant.html)
I would like to know how can I find the exact settings of the discriminant learner in the model generated by above code. Thank you very much!

  0 Comments

Sign in to comment.

Accepted Answer

Don Mathis
Don Mathis on 7 Nov 2018
Edited: Don Mathis on 7 Nov 2018
The property
classificationEnsemble.Trained
Contains the 30 individual Disciminant models that make up your ensemble. You can look at their parameters indviidually, for example:
classificationEnsemble.Trained{1}.Mu
Tip: you can call 'help' on an object to see all the properties and methods on it. In this case, it shows that there's a Trained property:
help classificationEnsemble

  0 Comments

Sign in to comment.

More Answers (0)

Sign in to answer this question.