does regression learner bagged tree is random forest or just bagged tree?

does anyone know the bagged tree in the regression learner app using algorithm of bagged tree or random forest ?

i have read that it using fitrensemble instead of treebagger code in the app, where treebagger code default is using the random forest algorithm.

Does it mean regression learner bagged tree is not random forest?

Answer 1

[Drew]

The short answer is that the "Bagged Trees" preset in the Regression Learner app does create a random forest of bagged trees using the fitrensenble function.

The page https://www.mathworks.com/help/stats/ensemble-algorithms.html contains details regarding:

(1) The meaning of "bagged trees" and "random forest". "Bootstrap aggregation (bagging) is a type of ensemble learning. To bag a weak learner such as a decision tree on a data set, generate many bootstrap replicas of the data set and grow decision trees on the replicas. Obtain each bootstrap replica by randomly selecting N out of N observations with replacement, where N is the data set size. In addition, every tree in the ensemble can randomly select predictors for each decision split, a technique called random forest [2] known to improve the accuracy of bagged trees." See the section https://www.mathworks.com/help/stats/ensemble-algorithms.html#bsw8at7 for more info.

(2) "TreeBagger and bagged ensembles (ClassificationBaggedEnsemble and RegressionBaggedEnsemble) share most functionalities, but not all. Additionally, some functionalities have different names." See the section https://www.mathworks.com/help/stats/ensemble-algorithms.html#bsxabwd for more info.