Can I use the fitcensemble function to create a Random Forest model?

I am new to machine learning classifiers.

I want to develop a Random Forest classification model with my IMU data.

I used the code below, and my question is: Can I call it a Random Forest model, or should I just call it a bagging ensemble model?

baggedEnsemble = fitcensemble(data_train(:,2:end), data_train(:,1),'Method','Bag','NumLearningCycles',100);

Answer 1

[Govind KM]

Hi Junwoo,

The primary characteristic of a Random Forest model is to use random subsets of predictors to train individual trees. In the “fitcensemble” function, if the “Method” parameter is passed as “Bag”, then by default the function uses bagging with decision trees as the base learners, utilizing random predictor selections at each split (Square root of the number of predictors by default). This makes it effectively the same as a Random Forest model.

To use bagging without the random selections, the “NumVariablesToSample” parameter can be set to “all”. The Ensemble aggregation method or the weak learners can also be changed as required to make the model created through the “fitcensemble” function different from a Random Forest model. You can refer to the documentation for more details on different options for the ensemble model, including ensemble method and weak learners:

https://www.mathworks.com/help/stats/fitcensemble.html#d126e410507

Hope this resoves the confusion.

[Junwoo]

Thank you so much!!

Your answer really helped me.