You can achieve this by using the predict function and the node property of the tree. Refer the steps below and the code snippet that uses synthetic data to depict the algorithm:
- Predict the leaf node for each data point using the predict function.
- Extract the unique leaf nodes from the predictions.
- Construct a logical matrix where each row corresponds to a data point and each column corresponds to a leaf.
% Generate synthetic data
rng(42); % Set random seed for reproducibility
n = 10; % Number of data points
numFeatures = 3; % Number of features
X = randn(n, numFeatures); % Random features
Y = X(:,1) * 2 + X(:,2) * -3 + X(:,3) * 1.5 + randn(n,1) * 0.5; % Target values with noise
% Train regression tree
Mdl = fitrtree(X, Y);
% Predict leaf assignments for each data point
[~, node] = predict(Mdl, X);
% Get unique leaf nodes
uniqueLeaves = unique(node);
tl = numel(uniqueLeaves);
n = numel(node);
% Initialize logical matrix
z = false(n, tl);
% Populate logical matrix
for i = 1:n
leafIndex = find(uniqueLeaves == node(i));
z(i, leafIndex) = true;
end
% Display the logical matrix (optional)
disp(z);
This method ensures that z(i,j) = 1 if data point i belongs to leaf j. Since it dynamically adapts to different datasets, it should work across various scenarios.
