How to define fitcensemble matlab function?

I am using following matlab code with latest MATLAB R2024b software

initIdx = 1:incrementSize;
Xinit = trainFeatures(initIdx,:);
Yinit = trainLabels(initIdx);
disp(size(Xinit));
disp(size(Yinit));

% Train initial batch model using fitcensemble
t = templateTree('MaxNumSplits',20); % You can adjust as needed
Mdl = fitcensemble(Xinit, Yinit, ...
    'Method', 'Bag', ...
    'NumLearningCycles', 100, ...
    'Learners', t, ...
    'OOBPrediction', 'on');

% Convert to incremental model
IncMdl = incrementalLearner(Mdl);

it is displaying [the following] error

3048 2

3048 1

Error using classreg.learning.FitTemplate/fillIfNeeded (line 734)

OOBPrediction is not a valid parameter name.

Error in classreg.learning.FitTemplate.make (line 140)

temp = fillIfNeeded(temp,type);

^^^^^^^^^^^^^^^^^^^^^^^

Error in fitensemble (line 363)

temp = classreg.learning.FitTemplate.make(method,'nlearn',nlearn,'learners',learners,varargin{:});

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Error in fitcensemble (line 215)

this = fitensemble(X, Y, Method, NumLearningCycles, Learners, ...

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Mdl = fitcensemble(Xinit, Yinit, ...

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

I request you to please suggest me how to resolve this issue.

[Abhaya]

Hi Aksh,

The error 'OOBPrediction is not a valid parameter name' indicates that MATLAB 'fitcensemble' method does not have any property named 'OOBPrediction'.

You can refer to the following MATLAB documentation on 'fitensenmle' method to check all the supported input arguments.

https://www.mathworks.com/help/releases/R2024b/stats/fitcensemble.html

Feel free to mention if you have any specific requirement.

[Walter Roberson]

OOBPrediction is a treebagger parameter

[Aksh Kumar]

Thank you very much for your suggestion. Now I am using

t = templateTree('MaxNumSplits',20);

Mdl= fitcensemble(Xinit, Yinit,'Method', 'Bag', 'NumLearningCycles', 100, 'Learners', t);

disp(Mdl);

ResponseName: 'Y'

CategoricalPredictors: []

ClassNames: [BARELAND BUILTUP BUSHES PLANT RATOON WATER]

ScoreTransform: 'none'

NumObservations: 3048

NumTrained: 100

Method: 'Bag'

LearnerNames: {'Tree'}

ReasonForTermination: 'Terminated normally after completing the requested number of training cycles.'

FitInfo: []

FitInfoDescription: 'None'

FResample: 1

Replace: 1

UseObsForLearner: [3048×100 logical]

% Convert to incremental model

IncMdl = incrementalLearner(Mdl);

This is giving error as Incorrect number or types of inputs or outputs for function incrementalLearner.

Please suggest me how to resolve this issue.

[Abhaya]

The error occurs because the MATLAB 'incrementalLearner' function expects a Linear regression model as an input.

The fitcensemble function, used to create ensemble classifiers, does not natively support incremental learning. This means that each time you call fitcensemble(), the ensemble model is rebuilt from scratch rather than adding new learners to an existing ensemble.

[Aksh Kumar]

Thanks a lot for your kind suggestion and help. May I request you to please provide me an alternative to given below code so that it works smoothely.

% Train initial batch model using fitcensemble
t = templateTree('MaxNumSplits',20); % You can adjust as needed
Mdl = fitcensemble(Xinit, Yinit, 'Method', 'Bag',...
    'NumLearningCycles', 100, 'Learners', t);
% Convert to incremental model
IncMdl = incrementalLearner(Mdl);

I would be highly obliged to you.

[Walter Roberson]

There is no way to do incremental learning on a fitcensemble model.

[Aksh Kumar]

Thank you very much for your guidence and help. I am trying to use following model

Mdl = TreeBagger(100, Xinit, Yinit, 'Method', 'classification', 'OOBPrediction', 'on');
% Convert to incremental model
IncMdl = incrementalClassificationEnsemble(Mdl);

but this giving error as

Unrecognized function or variable 'incrementalClassificationEnsemble'.

Please suggest me how fix this error

[Walter Roberson]

There is simply no function incrementalClassifictionEnsemble() .

There are functions incrementalDriftAwareLearner and incrementalClassificationKernel and incrementalClassificationLinear and incrementalClassificationECOC and incrementalClassificationNaiveBayes and incrementalConceptDriftDetector -- but there is no incrementalClassificationEnsemble()

There is no way to convert a classification ensemble to incremental learning.

[Aksh Kumar]

I would like to request you to please let me know whether this Incremental Classification is different than what you suggested that There is no way to convert a classification ensemble to incremental learning.

Looking forward for your kind suggestion.

[Walter Roberson]

TreeBagger() as a function returns a TreeBagger object https://www.mathworks.com/help/stats/treebagger.html

There are no functions to convert TreeBagger objects to incremental learning objects.

The documentation at https://www.mathworks.com/help/stats/incremental-learning-overview.html#mw_b9f908d5-71f7-419a-9f6c-386f264864b9 describes the objects that can be converted to incremental learning.

• ClassificationECOC and CompactClassificationECOC
• ClassificationKernel
• ClassificationSVM and CompactClassificationSVM
• ClassificationLinear
• ClassificationNaiveBayes
• RegressionKernel
• RegressionSVM and CompactRegressionSVM
• RegressionLinear

Notice that TreeBagger() and ClassificationBaggedEnsemble and ClassificationEnsemble and ClassificationPartitionedEnsemble are not on this list.