bat365 Homepage

ClassificationEnsemble

Package: classreg.learning.classif
Superclasses: CompactClassificationEnsemble

Ensemble classifier

Description

ClassificationEnsemble combines a set of trained weak learner models and data on which these learners were trained. It can predict ensemble response for new data by aggregating predictions from its weak learners. It stores data used for training, can compute resubstitution predictions, and can resume training if desired.

Construction

Create a classification ensemble object (ens) using fitcensemble.

Properties

`BinEdges`	Bin edges for numeric predictors, specified as a cell array of p numeric vectors, where p is the number of predictors. Each vector includes the bin edges for a numeric predictor. The element in the cell array for a categorical predictor is empty because the software does not bin categorical predictors. The software bins numeric predictors only if you specify the `'NumBins'` name-value argument as a positive integer scalar when training a model with tree learners. The `BinEdges` property is empty if the `'NumBins'` value is empty (default). You can reproduce the binned predictor data `Xbinned` by using the `BinEdges` property of the trained model `mdl`. X = mdl.X; % Predictor data Xbinned = zeros(size(X)); edges = mdl.BinEdges; % Find indices of binned predictors. idxNumeric = find(~cellfun(@isempty,edges)); if iscolumn(idxNumeric) idxNumeric = idxNumeric'; end for j = idxNumeric x = X(:,j); % Convert x to array if x is a table. if istable(x) x = table2array(x); end % Group x into bins by using the `discretize` function. xbinned = discretize(x,[-inf; edges{j}; inf]); Xbinned(:,j) = xbinned; end `Xbinned` contains the bin indices, ranging from 1 to the number of bins, for numeric predictors. `Xbinned` values are 0 for categorical predictors. If `X` contains `NaN`s, then the corresponding `Xbinned` values are `NaN`s.
`CategoricalPredictors`	Categorical predictor indices, specified as a vector of positive integers. `CategoricalPredictors` contains index values indicating that the corresponding predictors are categorical. The index values are between 1 and `p`, where `p` is the number of predictors used to train the model. If none of the predictors are categorical, then this property is empty (`[]`).
`ClassNames`	List of the elements in `Y` with duplicates removed. `ClassNames` can be a numeric vector, categorical vector, logical vector, character array, or cell array of character vectors. `ClassNames` has the same data type as the data in the argument `Y`. (The software treats string arrays as cell arrays of character vectors.)
`CombineWeights`	Character vector describing how `ens` combines weak learner weights, either `'WeightedSum'` or `'WeightedAverage'`.
`Cost`	Square matrix, where `Cost(i,j)` is the cost of classifying a point into class `j` if its true class is `i` (the rows correspond to the true class and the columns correspond to the predicted class). The order of the rows and columns of `Cost` corresponds to the order of the classes in `ClassNames`. The number of rows and columns in `Cost` is the number of unique classes in the response. This property is read-only.
`ExpandedPredictorNames`	Expanded predictor names, stored as a cell array of character vectors. If the model uses encoding for categorical variables, then `ExpandedPredictorNames` includes the names that describe the expanded variables. Otherwise, `ExpandedPredictorNames` is the same as `PredictorNames`.
`FitInfo`	Numeric array of fit information. The `FitInfoDescription` property describes the content of this array.
`FitInfoDescription`	Character vector describing the meaning of the `FitInfo` array.
`HyperparameterOptimizationResults`	Description of the cross-validation optimization of hyperparameters, stored as a `BayesianOptimization` object or a table of hyperparameters and associated values. Nonempty when the `OptimizeHyperparameters` name-value pair is nonempty at creation. Value depends on the setting of the `HyperparameterOptimizationOptions` name-value pair at creation: `'bayesopt'` (default) — Object of class `BayesianOptimization` `'gridsearch'` or `'randomsearch'` — Table of hyperparameters used, observed objective function values (cross-validation loss), and rank of observations from lowest (best) to highest (worst)
`LearnerNames`	Cell array of character vectors with names of weak learners in the ensemble. The name of each learner appears just once. For example, if you have an ensemble of 100 trees, `LearnerNames` is `{'Tree'}`.
`Method`	Character vector describing the method that creates `ens`.
`ModelParameters`	Parameters used in training `ens`.
`NumObservations`	Numeric scalar containing the number of observations in the training data.
`NumTrained`	Number of trained weak learners in `ens`, a scalar.
`PredictorNames`	Cell array of names for the predictor variables, in the order in which they appear in `X`.
`Prior`	Numeric vector of prior probabilities for each class. The order of the elements of `Prior` corresponds to the order of the classes in `ClassNames`. The number of elements of `Prior` is the number of unique classes in the response. This property is read-only.
`ReasonForTermination`	Character vector describing the reason `fitcensemble` stopped adding weak learners to the ensemble.
`ResponseName`	Character vector with the name of the response variable `Y`.
`RowsUsed`	Rows of the original training data stored in the model, specified as a logical vector. This property is empty if all rows are stored in `X` and `Y`.
`ScoreTransform`	Function handle for transforming scores, or character vector representing a built-in transformation function. `'none'` means no transformation; equivalently, `'none'` means `@(x)x`. For a list of built-in transformation functions and the syntax of custom transformation functions, see `fitctree`. Add or change a `ScoreTransform` function using dot notation: ens.ScoreTransform = 'function' or ens.ScoreTransform = @function
`Trained`	A cell vector of trained classification models. If `Method` is `'LogitBoost'` or `'GentleBoost'`, then `ClassificationEnsemble` stores trained learner `j` in the `CompactRegressionLearner` property of the object stored in `Trained{j}`. That is, to access trained learner `j`, use `ens.Trained{j}.CompactRegressionLearner`. Otherwise, cells of the cell vector contain the corresponding, compact classification models.
`TrainedWeights`	Numeric vector of trained weights for the weak learners in `ens`. `TrainedWeights` has `T` elements, where `T` is the number of weak learners in `learners`.
`UsePredForLearner`	Logical matrix of size `P`-by-`NumTrained`, where `P` is the number of predictors (columns) in the training data `X`. `UsePredForLearner(i,j)` is `true` when learner `j` uses predictor `i`, and is `false` otherwise. For each learner, the predictors have the same order as the columns in the training data `X`. If the ensemble is not of type `Subspace`, all entries in `UsePredForLearner` are `true`.
`W`	Scaled `weights`, a vector with length `n`, the number of rows in `X`. The sum of the elements of `W` is `1`.
`X`	Matrix or table of predictor values that trained the ensemble. Each column of `X` represents one variable, and each row represents one observation.
`Y`	Numeric vector, categorical vector, logical vector, character array, or cell array of character vectors. Each row of `Y` represents the classification of the corresponding row of `X`.

Object Functions

`compact`	Reduce size of classification ensemble model
`compareHoldout`	Compare accuracies of two classification models using new data
`crossval`	Cross-validate classification ensemble model
`edge`	Classification edge for classification ensemble model
`gather`	Gather properties of Statistics and Machine Learning Toolbox object from GPU
`lime`	Local interpretable model-agnostic explanations (LIME)
`loss`	Classification loss for classification ensemble model
`margin`	Classification margins for classification ensemble model
`partialDependence`	Compute partial dependence
`plotPartialDependence`	Create partial dependence plot (PDP) and individual conditional expectation (ICE) plots
`predict`	Classify observations using ensemble of classification models
`predictorImportance`	Estimates of predictor importance for classification ensemble of decision trees
`resubEdge`	Resubstitution classification edge for classification ensemble model
`resubLoss`	Resubstitution classification loss for classification ensemble model
`resubMargin`	Resubstitution classification margins for classification ensemble model
`resubPredict`	Classify observations in classification ensemble model
`resume`	Resume training of classification ensemble model
`shapley`	Shapley values
`testckfold`	Compare accuracies of two classification models by repeated cross-validation

Copy Semantics

Value. To learn how value classes affect copy operations, see Copying Objects.

Examples

collapse all

Train Boosted Classification Ensemble

Open Live Script

Load the ionosphere data set.

load ionosphere

Train a boosted ensemble of 100 classification trees using all measurements and the AdaBoostM1 method.

Mdl = fitcensemble(X,Y,'Method','AdaBoostM1')

Mdl = 
  ClassificationEnsemble
             ResponseName: 'Y'
    CategoricalPredictors: []
               ClassNames: {'b'  'g'}
           ScoreTransform: 'none'
          NumObservations: 351
               NumTrained: 100
                   Method: 'AdaBoostM1'
             LearnerNames: {'Tree'}
     ReasonForTermination: 'Terminated normally after completing the requested number of training cycles.'
                  FitInfo: [100x1 double]
       FitInfoDescription: {2x1 cell}

Mdl is a ClassificationEnsemble model object.

Mdl.Trained is the property that stores a 100-by-1 cell vector of the trained classification trees (CompactClassificationTree model objects) that compose the ensemble.

Plot a graph of the first trained classification tree.

view(Mdl.Trained{1},'Mode','graph')

Figure Classification tree viewer contains an axes object and other objects of type uimenu, uicontrol. The axes object contains 36 objects of type line, text. One or more of the lines displays its values using only markers

By default, fitcensemble grows shallow trees for boosted ensembles of trees.

Predict the label of the mean of X.

predMeanX = predict(Mdl,mean(X))

predMeanX = 1x1 cell array
    {'g'}

Tips

For an ensemble of classification trees, the Trained property of ens stores an ens.NumTrained-by-1 cell vector of compact classification models. For a textual or graphical display of tree t in the cell vector, enter:

view(ens.Trained{t}.CompactRegressionLearner) for ensembles aggregated using LogitBoost or GentleBoost.
view(ens.Trained{t}) for all other aggregation methods.

Extended Capabilities

C/C++ Code Generation
Generate C and C++ code using MATLAB® Coder™.

Usage notes and limitations:

The predict function supports code generation.
To integrate the prediction of an ensemble into Simulink^®, you can use the ClassificationEnsemble Predict block in the Statistics and Machine Learning Toolbox™ library or a MATLAB^® Function block with the predict function.
When you train an ensemble by using fitcensemble, the following restrictions apply.
- The value of the ScoreTransform name-value argument cannot be an anonymous function.
- Code generation limitations for the weak learners used in the ensemble also apply to the ensemble.
  - For decision tree weak learners, you cannot use surrogate splits; that is, the value of the Surrogate name-value argument must be 'off'.
  - For k-nearest neighbor weak learners, the value of the Distance name-value argument cannot be a custom distance function. The value of the DistanceWeight name-value argument can be a custom distance weight function, but it cannot be an anonymous function.
For fixed-point code generation, the following additional restrictions apply.
- When you train an ensemble by using fitcensemble, you must train an ensemble using tree learners, and the ScoreTransform value cannot be 'invlogit'.
- Categorical predictors (logical, categorical, char, string, or cell) are not supported. You cannot use the CategoricalPredictors name-value argument. To include categorical predictors in a model, preprocess them by using dummyvar before fitting the model.
- Class labels with the categorical data type are not supported. Both the class label value in the training data (Tbl or Y) and the value of the ClassNames name-value argument cannot be an array with the categorical data type.

For more information, see Introduction to Code Generation.

GPU Arrays
Accelerate code by running on a graphics processing unit (GPU) using Parallel Computing Toolbox™.

Usage notes and limitations:

The following object functions fully support GPU arrays:
The following object functions offer limited support for GPU arrays:
The object functions execute on a GPU if either of the following apply:
- The model was fitted with GPU arrays.
- The predictor data that you pass to the object function is a GPU array.

For more information, see Run MATLAB Functions on a GPU (Parallel Computing Toolbox).

Version History

Introduced in R2011a

expand all

R2023b: Model with discriminant analysis weak learners stores observations with missing predictor values

Starting in R2023b, training observations with missing predictor values are included in the X, Y, and W data properties of classification ensemble models with discriminant analysis weak learners. The RowsUsed property indicates the training observations stored in the model, rather than those used for training. Observations with missing predictor values continue to be omitted from the model training process.

In previous releases, the software omitted training observations that contained missing predictor values from the data properties of the model.

R2022a: `Cost` property stores the user-specified cost matrix

Starting in R2022a, the Cost property stores the user-specified cost matrix, so that you can compute the observed misclassification cost using the specified cost value. The software stores normalized prior probabilities (Prior) and observation weights (W) that do not reflect the penalties described in the cost matrix. To compute the observed misclassification cost, specify the LossFun name-value argument as "classifcost" when you call the loss or resubLoss function.

Note that model training has not changed and, therefore, the decision boundaries between classes have not changed.

For training, the fitting function updates the specified prior probabilities by incorporating the penalties described in the specified cost matrix, and then normalizes the prior probabilities and observation weights. This behavior has not changed. In previous releases, the software stored the default cost matrix in the Cost property and stored the prior probabilities and observation weights used for training in the Prior and W properties, respectively. Starting in R2022a, the software stores the user-specified cost matrix without modification, and stores normalized prior probabilities and observation weights that do not reflect the cost penalties. For more details, see Misclassification Cost Matrix, Prior Probabilities, and Observation Weights.

Some object functions use the Cost, Prior, and W properties:

The loss and resubLoss functions use the cost matrix stored in the Cost property if you specify the LossFun name-value argument as "classifcost" or "mincost".
The loss and edge functions use the prior probabilities stored in the Prior property to normalize the observation weights of the input data.
The resubLoss and resubEdge functions use the observation weights stored in the W property.

If you specify a nondefault cost matrix when you train a classification model, the object functions return a different value compared to previous releases.

If you want the software to handle the cost matrix, prior probabilities, and observation weights in the same way as in previous releases, adjust the prior probabilities and observation weights for the nondefault cost matrix, as described in Adjust Prior Probabilities and Observation Weights for Misclassification Cost Matrix. Then, when you train a classification model, specify the adjusted prior probabilities and observation weights by using the Prior and Weights name-value arguments, respectively, and use the default cost matrix.