MATLAB: Get regions from treebagger

random forestStatistics and Machine Learning Toolboxtreebagger

Hi,

I have a random forest, trained through treeBagger function. I need to check for the new point x, whether it results in a same region R_t for all t, with historical point x_j, for all j. Please let me know me how I can get it.

Best, Afshin

Best Answer

Thanks for the clarification!

Let me assume that you've trained an ensemble of trees using the TreeBagger function, and that this bag of trees is called 'B'.

A cell vector of the trained CompactRegressionTree subtrees is stored in the ‘Trees’ property of B. Assuming that you’ve extracted one of these trees, ‘tree1’, as follows,

tree1 = B.Trees{1};

You can determine the leaf node that determines the predicted output for the observation (or array of observations) X, using the 'predict' function:

[~,node] = predict(tree1,X)

https://www.mathworks.com/help/stats/compactregressiontree.predict.html

If you would like the parent of the leaf node (e.g, the node that ‘splits’ the data), you can look up the node’s parent in the ‘Parent’ property of ‘tree1’:

splittingNode = tree1.Parent(node);

Related Solutions

MATLAB: How to report Terminal Node (only the one use to predict Y) Standard Deviation in TreeBagger

TreeBagger is an ensemble of trees. All trees are different because they were grown on different bootstrap replicas of the data. For example, one tree can have say 50 nodes and another one can have 60 nodes. Even if both trees have the same number of nodes, the tree structures could be very different. So what is "terminal nodes standard deviation average in TreeBagger"? You could, I suppose, take all leaves (terminal nodes) from all trees in the ensemble and compute the standard deviation over them. I have never seen anyone do this. What statistical interpretation would this number have?

If you want to obtain the standard deviation for a node in a single regression tree, you could use nodeerr method of classregtree. For example

load carsmall
X = [Acceleration Displacement Horsepower Weight];
b = TreeBagger(10,X,MPG,'method','reg')
nodeerr(b.Trees{1})

would give you node variance (standard error squared) for the 1st tree in the ensemble. This number has a clear interpretation: It is the uncertainty of prediction for an observation landing on this node of the tree.

MATLAB: How to select number of trees and leaf size in bagged regression trees

Hi,

Whenever you are dealing with machine learning models there is no specific rule to take the parameters, these parameters change according to dataset and the result you are expecting.

To answer your question if processing time and memory is not a constraint in your case you can use as many trees as possible. In general, the more trees you use the better results you get.

When it comes to the number of leaf nodes, you don’t want your model to overfit. Use Bias vs Variance trade-off in order to choose the number of leaf nodes wrt your dataset.

For implementation level information refer to the links:

Best Answer

Related Solutions

MATLAB: How to report Terminal Node (only the one use to predict Y) Standard Deviation in TreeBagger

MATLAB: How to select number of trees and leaf size in bagged regression trees

Related Question