CompactRegressionEnsemble
statistics: CompactRegressionEnsemble
Compact ensemble of regression trees
A CompactRegressionEnsemble object carries the trained trees of a
boosted or bagged regression ensemble and what prediction needs, but not
the observations it was fitted on. It predicts new data identically to
the ensemble it came from, and trees can be removed from it.
Create one with the compact method of a
RegressionEnsemble or RegressionBaggedEnsemble object.
See also: fitrensemble, RegressionEnsemble, RegressionBaggedEnsemble
Source Code: CompactRegressionEnsemble
The CompactRegressionEnsemble class contains the following properties:
A cell array of character vectors. This property is read-only.
The predictors every tree treats as categorical, empty when none is. This property is read-only.
This property is read-only.
The same as PredictorNames. This property is read-only.
Always empty, as MATLAB returns it for tree learners. This property is read-only.
This property is read-only.
A column cell array of CompactRegressionTree objects. This
property is read-only.
A column with one weight per tree. This property is read-only.
'WeightedSum' for LSBoost, whose prediction is the sum of each
tree’s prediction times its weight, or 'WeightedAverage' for
Bag, whose prediction is the weighted average of its trees’. This
property is read-only.
'none' (default), 'exp', 'log' or a function
handle. Predictions and losses use the transformed response.
The CompactRegressionEnsemble class offers the following public methods:
CompactRegressionEnsemble: yfit = predict (obj, X)
CompactRegressionEnsemble: yfit = predict (…, name, value)
yfit holds, for each row of X, the sum over the trees of
each tree’s prediction times its weight for LSBoost, or the weighted
average of the trees’ predictions for Bag, after
ResponseTransform. A row that no tree may predict is
NaN.
Name-Value arguments:
| Name | Value | |
|---|---|---|
'Learners' | A vector of indices of the trees to use. The default is all of them. | |
'UseObsForLearner' | An NxNumTrained logical matrix saying which tree may predict which row. The default lets every tree predict every row. |
See also: CompactRegressionEnsemble, fitrensemble
CompactRegressionEnsemble: L = loss (obj, X, Y)
CompactRegressionEnsemble: L = loss (…, name, value)
L is the weighted mean squared error of the predictions for the rows of X against Y, the weights normalized to sum to one over the rows that have a response; a row whose response is missing is left out.
Name-Value arguments:
| Name | Value | |
|---|---|---|
'LossFun' | 'mse' (default) or a
function handle called as lossfun (Y, Yfit, W), with column
vectors of the responses, the predictions and the normalized weights,
returning a scalar. | |
'Mode' | 'ensemble' (default) for one
loss over the trees used, 'cumulative' for a column whose
element j uses the first j of them, or
'individual' for a column with the loss of each on its own. | |
'Weights' | A nonnegative vector with one weight per row. The default is uniform. |
'Learners' and 'UseObsForLearner' are taken as by
predict. A row that no tree may predict is left out and the
weights are renormalized over the rest.
See also: CompactRegressionEnsemble, CompactRegressionEnsemble.predict
CompactRegressionEnsemble: imp = predictorImportance (obj)
CompactRegressionEnsemble: [imp, ma] = predictorImportance (obj)
imp is a row vector with one element per predictor, the average
over the trees of each tree’s predictorImportance, weighted by
TrainedWeights. ma, the predictive measure of
association between the predictors, is empty, the trees growing no
surrogate splits.
See also: CompactRegressionEnsemble
CompactRegressionEnsemble: C = removeLearners (obj, idx)
C is the ensemble without the trees whose indices idx holds, their weights removed with them.
See also: CompactRegressionEnsemble