You can use Classification Learner to automatically train a selection of different classification models on your data. Use automated training to quickly try a selection of model types, then explore promising models interactively. To get started, try these options first:
Get Started Classifier Buttons  Description 

All QuickToTrain  Try this first. The app will train all the model types available for your data set that are typically fast to fit. 
All Linear  Try this if you expect linear boundaries between the classes in your data. This option fits only Linear SVM and Linear Discriminant. 
All  Use this to train all available nonoptimizable model types. Trains every type regardless of any prior trained models. Can be timeconsuming. 
See Automated Classifier Training.
If you want to explore classifiers one at a time, or you already know what classifier type you want, you can select individual models or train a group of the same type. To see all available classifier options, on the Classification Learner tab, click the arrow in the Model Type section to expand the list of classifiers. The nonoptimizable model options in the Model Type gallery are preset starting points with different settings, suitable for a range of different classification problems. To use optimizable model options and tune model hyperparameters automatically, see Hyperparameter Optimization in Classification Learner App.
For help choosing the best classifier type for your problem, see the table showing typical characteristics of different supervised learning algorithms. Use the table as a guide for your final choice of algorithms. Decide on the tradeoff you want in speed, memory usage, flexibility, and interpretability. The best classifier type depends on your data.
Tip
To avoid overfitting, look for a model of lower flexibility that provides sufficient accuracy. For example, look for simple models such as decision trees and discriminants that are fast and easy to interpret. If the models are not accurate enough predicting the response, choose other classifiers with higher flexibility, such as ensembles. To control flexibility, see the details for each classifier type.
Characteristics of Classifier Types
Classifier  Prediction Speed  Memory Usage  Interpretability 

Decision Trees
 Fast  Small  Easy 
Discriminant Analysis
 Fast  Small for linear, large for quadratic  Easy 
Logistic Regression
 Fast  Medium  Easy 
Naive Bayes Classifiers
 Medium for simple distributions Slow for kernel distributions or highdimensional data  Small for simple distributions Medium for kernel distributions or highdimensional data  Easy 
Medium for linear Slow for others  Medium for linear. All others: medium for multiclass, large for binary.  Easy for Linear SVM. Hard for all other kernel types.  
Nearest Neighbor Classifiers
 Slow for cubic Medium for others  Medium  Hard 
Ensemble Classifiers
 Fast to medium depending on choice of algorithm  Low to high depending on choice of algorithm  Hard 
Neural Network Classifiers
 Fast  Medium  Hard 
The tables on this page describe general characteristics of speed and memory usage for all the nonoptimizable preset classifiers. The classifiers were tested with various data sets (up to 7000 observations, 80 predictors, and 50 classes), and the results define the following groups:
Speed
Fast 0.01 second
Medium 1 second
Slow 100 seconds
Memory
Small 1MB
Medium 4MB
Large 100MB
These tables provide a general guide. Your results depend on your data and the speed of your machine.
To read a description of each classifier in Classification Learner, switch to the details view.
Tip
After you choose a classifier type (for example, decision trees), try training using each of the classifiers. The nonoptimizable options in the Model Type gallery are starting points with different settings. Try them all to see which option produces the best model with your data.
For workflow instructions, see Train Classification Models in Classification Learner App.
In Classification Learner, the Model Type gallery shows as available the classifier types that support your selected data.
Classifier  All predictors numeric  All predictors categorical  Some categorical, some numeric 

Decision Trees  Yes  Yes  Yes 
Discriminant Analysis  Yes  No  No 
Logistic Regression  Yes  Yes  Yes 
Naive Bayes  Yes  Yes  Yes 
SVM  Yes  Yes  Yes 
Nearest Neighbor  Euclidean distance only  Hamming distance only  No 
Ensembles  Yes  Yes, except Subspace Discriminant  Yes, except any Subspace 
Neural Networks  Yes  Yes  Yes 
Decision trees are easy to interpret, fast for fitting and prediction, and low on memory usage, but they can have low predictive accuracy. Try to grow simpler trees to prevent overfitting. Control the depth with the Maximum number of splits setting.
Tip
Model flexibility increases with the Maximum number of splits setting.
Classifier Type  Prediction Speed  Memory Usage  Interpretability  Model Flexibility 

Coarse Tree  Fast  Small  Easy  Low Few leaves to make coarse distinctions between classes (maximum number of splits is 4). 
Medium Tree  Fast  Small  Easy  Medium Medium number of leaves for finer distinctions between classes (maximum number of splits is 20). 
Fine Tree  Fast  Small  Easy  High Many leaves to make many fine distinctions between classes (maximum number of splits is 100). 
Tip
In the Model Type gallery click All Trees to try each of the nonoptimizable decision tree options. Train them all to see which settings produce the best model with your data. Select the best model in the Models pane. To try to improve your model, try feature selection, and then try changing some advanced options.
You train classification trees to predict responses to data. To predict a response, follow the decisions in the tree from the root (beginning) node down to a leaf node. The leaf node contains the response. Statistics and Machine Learning Toolbox™ trees are binary. Each step in a prediction involves checking the value of one predictor (variable). For example, here is a simple classification tree:
This tree predicts classifications based on two predictors, x1
and x2
. To predict, start at the top node. At each decision,
check the values of the predictors to decide which branch to follow. When the
branches reach a leaf node, the data is classified either as type
0
or 1
.
You can visualize your decision tree model by exporting the model from the app, and then entering:
view(trainedModel.ClassificationTree,'Mode','graph')
fisheriris
data.
Tip
For an example, see Train Decision Trees Using Classification Learner App.
Classification trees in Classification Learner use the fitctree
function. You can set
these options:
Maximum number of splits
Specify the maximum number of splits or branch points to control the depth of your tree. When you grow a decision tree, consider its simplicity and predictive power. To change the number of splits, click the buttons or enter a positive integer value in the Maximum number of splits box.
A fine tree with many leaves is usually highly accurate on the training data. However, the tree might not show comparable accuracy on an independent test set. A leafy tree tends to overtrain, and its validation accuracy is often far lower than its training (or resubstitution) accuracy.
In contrast, a coarse tree does not attain high training accuracy. But a coarse tree can be more robust in that its training accuracy can approach that of a representative test set. Also, a coarse tree is easy to interpret.
Split criterion
Specify the split criterion measure for deciding when to split nodes. Try each of the three settings to see if they improve the model with your data.
Split criterion options are Gini's diversity
index
, Twoing rule
, or
Maximum deviance reduction
(also known as
cross entropy).
The classification tree tries to optimize to pure nodes containing only one class. Gini's diversity index (the default) and the deviance criterion measure node impurity. The twoing rule is a different measure for deciding how to split a node, where maximizing the twoing rule expression increases node purity.
For details of these split criteria, see
ClassificationTree
More About.
Surrogate decision splits — Only for missing data.
Specify surrogate use for decision splits. If you have data with missing values, use surrogate splits to improve the accuracy of predictions.
When you set Surrogate decision splits to
On
, the classification tree finds at most
10 surrogate splits at each branch node. To change the number, click the
buttons or enter a positive integer value in the Maximum
surrogates per node box.
When you set Surrogate decision splits to
Find All
, the classification tree finds
all surrogate splits at each branch node. The Find
All
setting can use considerable time and
memory.
Alternatively, you can let the app choose some of these model options automatically by using hyperparameter optimization. See Hyperparameter Optimization in Classification Learner App.
Discriminant analysis is a popular first classification algorithm to try because it is fast, accurate and easy to interpret. Discriminant analysis is good for wide datasets.
Discriminant analysis assumes that different classes generate data based on different Gaussian distributions. To train a classifier, the fitting function estimates the parameters of a Gaussian distribution for each class.
Classifier Type  Prediction Speed  Memory Usage  Interpretability  Model Flexibility 

Linear Discriminant  Fast  Small  Easy  Low Creates linear boundaries between classes. 
Quadratic
Discriminant  Fast  Large  Easy  Low Creates nonlinear boundaries between classes (ellipse, parabola or hyperbola). 
Discriminant analysis in Classification Learner uses the fitcdiscr
function. For both
linear and quadratic discriminants, you can change the Covariance
structure option. If you have predictors with zero variance or if
any of the covariance matrices of your predictors are singular, training can
fail using the default, Full
covariance structure. If
training fails, select the Diagonal
covariance
structure instead.
Alternatively, you can let the app choose some of these model options automatically by using hyperparameter optimization. See Hyperparameter Optimization in Classification Learner App.
If you have 2 classes, logistic regression is a popular simple classification algorithm to try because it is easy to interpret. The classifier models the class probabilities as a function of the linear combination of predictors.
Classifier Type  Prediction Speed  Memory Usage  Interpretability  Model Flexibility 

Logistic Regression  Fast  Medium  Easy  Low You cannot change any parameters to control model flexibility. 
Logistic regression in Classification Learner uses the fitglm
function. You cannot set any options for this classifier in
the app.
Naive Bayes classifiers are easy to interpret and useful for multiclass classification. The naive Bayes algorithm leverages Bayes theorem and makes the assumption that predictors are conditionally independent, given the class. Use these classifiers if this independence assumption is valid for predictors in your data. However, the algorithm still appears to work well when the independence assumption is not valid.
For kernel naive Bayes classifiers, you can control the kernel smoother type with the Kernel Type setting, and control the kernel smoothing density support with the Support setting.
Classifier Type  Prediction Speed  Memory Usage  Interpretability  Model Flexibility 

Gaussian Naive Bayes  Medium Slow for highdimensional data  Small Medium for highdimensional data  Easy  Low You cannot change any parameters to control model flexibility. 
Kernel Naive Bayes  Slow  Medium  Easy  Medium You can change settings for Kernel Type and Support to control how the classifier models predictor distributions. 
Naive Bayes in Classification Learner uses the fitcnb
function.
For kernel naive Bayes classifiers, you can set these options:
Kernel Type — Specify the kernel smoother type. Try setting each of these options to see if they improve the model with your data.
Kernel type options are Gaussian
,
Box
,
Epanechnikov
, or
Triangle
.
Support — Specify the kernel smoothing density support. Try setting each of these options to see if they improve the model with your data.
Support options are Unbounded
(all real
values) or Positive
(all positive real
values).
Alternatively, you can let the app choose some of these model options automatically by using hyperparameter optimization. See Hyperparameter Optimization in Classification Learner App.
For next steps training models, see Train Classification Models in Classification Learner App.
In Classification Learner, you can train SVMs when your data has two or more classes.
Classifier Type  Prediction Speed  Memory Usage  Interpretability  Model Flexibility 

Linear SVM  Binary: Fast Multiclass: Medium  Medium  Easy  Low Makes a simple linear separation between classes. 
Quadratic SVM  Binary: Fast Multiclass: Slow  Binary: Medium Multiclass: Large  Hard  Medium 
Cubic SVM  Binary: Fast Multiclass: Slow  Binary: Medium Multiclass: Large  Hard  Medium 
Fine Gaussian SVM  Binary: Fast Multiclass: Slow  Binary: Medium Multiclass: Large  Hard  High — decreases with kernel scale
setting. Makes finely detailed distinctions between classes, with kernel scale set to sqrt(P)/4 . 
Medium Gaussian SVM  Binary: Fast Multiclass: Slow  Binary: Medium Multiclass: Large  Hard  Medium Medium distinctions, with kernel scale set to sqrt(P) . 
Coarse Gaussian SVM  Binary: Fast Multiclass: Slow  Binary: Medium Multiclass: Large  Hard  Low Makes coarse distinctions between classes, with kernel scale set to sqrt(P)*4 , where P is
the number of predictors. 
Tip
Try training each of the nonoptimizable support vector machine options in the Model Type gallery. Train them all to see which settings produce the best model with your data. Select the best model in the Models pane. To try to improve your model, try feature selection, and then try changing some advanced options.
An SVM classifies data by finding the best hyperplane that separates data points of one class from those of the other class. The best hyperplane for an SVM means the one with the largest margin between the two classes. Margin means the maximal width of the slab parallel to the hyperplane that has no interior data points.
The support vectors are the data points that are closest to the separating hyperplane; these points are on the boundary of the slab. The following figure illustrates these definitions, with + indicating data points of type 1, and – indicating data points of type –1.
SVMs can also use a soft margin, meaning a hyperplane that separates many, but not all data points.
For an example, see Train Support Vector Machines Using Classification Learner App.
If you have exactly two classes, Classification Learner uses the fitcsvm
function to train the
classifier. If you have more than two classes, the app uses the fitcecoc
function to reduce the
multiclass classification problem to a set of binary classification subproblems,
with one SVM learner for each subproblem. To examine the code for the binary and
multiclass classifier types, you can generate code from your trained classifiers
in the app.
You can set these options in the app:
Kernel function
Specify the Kernel function to compute the classifier.
Linear kernel, easiest to interpret
Gaussian or Radial Basis Function (RBF) kernel
Quadratic
Cubic
Box constraint level
Specify the box constraint to keep the allowable values of the Lagrange multipliers in a box, a bounded region.
To tune your SVM classifier, try increasing the box constraint level. Click the buttons or enter a positive scalar value in the Box constraint level box. Increasing the box constraint level can decrease the number of support vectors, but also can increase training time.
The Box Constraint parameter is the softmargin penalty known as C in the primal equations, and is a hard “box” constraint in the dual equations.
Kernel scale mode
Specify manual kernel scaling if desired.
When you set Kernel scale mode to
Auto
, then the software uses a heuristic
procedure to select the scale value. The heuristic procedure uses
subsampling. Therefore, to reproduce results, set a random number seed
using rng
before training the classifier.
When you set Kernel scale mode to
Manual
, you can specify a value. Click
the buttons or enter a positive scalar value in the Manual
kernel scale box. The software divides all elements of
the predictor matrix by the value of the kernel scale. Then, the
software applies the appropriate kernel norm to compute the Gram
matrix.
Tip
Model flexibility decreases with the kernel scale setting.
Multiclass method
Only for data with 3 or more classes. This method reduces the
multiclass classification problem to a set of binary classification
subproblems, with one SVM learner for each subproblem.
OnevsOne
trains one learner for each
pair of classes. It learns to distinguish one class from the other.
OnevsAll
trains one learner for each
class. It learns to distinguish one class from all others.
Standardize data
Specify whether to scale each coordinate distance. If predictors have widely different scales, standardizing can improve the fit.
Alternatively, you can let the app choose some of these model options automatically by using hyperparameter optimization. See Hyperparameter Optimization in Classification Learner App.
Nearest neighbor classifiers typically have good predictive accuracy in low dimensions, but might not in high dimensions. They have high memory usage, and are not easy to interpret.
Tip
Model flexibility decreases with the Number of neighbors setting.
Classifier Type  Prediction Speed  Memory Usage  Interpretability  Model Flexibility 

Fine KNN  Medium  Medium  Hard  Finely detailed distinctions between classes. The number of neighbors is set to 1. 
Medium KNN  Medium  Medium  Hard  Medium distinctions between classes. The number of neighbors is set to 10. 
Coarse KNN  Medium  Medium  Hard  Coarse distinctions between classes. The number of neighbors is set to 100. 
Cosine KNN  Medium  Medium  Hard  Medium distinctions between classes, using a Cosine distance metric. The number of neighbors is set to 10. 
Cubic KNN  Slow  Medium  Hard  Medium distinctions between classes, using a cubic distance metric. The number of neighbors is set to 10. 
Weighted KNN  Medium  Medium  Hard  Medium distinctions between classes, using a distance weight. The number of neighbors is set to 10. 
Tip
Try training each of the nonoptimizable nearest neighbor options in the Model Type gallery. Train them all to see which settings produce the best model with your data. Select the best model in the Models pane. To try to improve your model, try feature selection, and then (optionally) try changing some advanced options.
What is kNearest Neighbor classification? Categorizing query points based on their distance to points (or neighbors) in a training dataset can be a simple yet effective way of classifying new points. You can use various metrics to determine the distance. Given a set X of n points and a distance function, knearest neighbor (kNN) search lets you find the k closest points in X to a query point or set of points. kNNbased algorithms are widely used as benchmark machine learning rules.
For an example, see Train Nearest Neighbor Classifiers Using Classification Learner App.
Nearest Neighbor classifiers in Classification Learner use the fitcknn
function. You can set these
options:
Number of neighbors
Specify the number of nearest neighbors to find for classifying each point when predicting. Specify a fine (low number) or coarse classifier (high number) by changing the number of neighbors. For example, a fine KNN uses one neighbor, and a coarse KNN uses 100. Many neighbors can be time consuming to fit.
Distance metric
You can use various metrics to determine the distance to points. For
definitions, see the class
ClassificationKNN
.
Distance weight
Specify the distance weighting function. You can choose
Equal
(no weights),
Inverse
(weight is 1/distance), or
Squared Inverse
(weight is
1/distance^{2}).
Standardize data
Specify whether to scale each coordinate distance. If predictors have widely different scales, standardizing can improve the fit.
Alternatively, you can let the app choose some of these model options automatically by using hyperparameter optimization. See Hyperparameter Optimization in Classification Learner App.
Ensemble classifiers meld results from many weak learners into one highquality ensemble model. Qualities depend on the choice of algorithm.
Tip
Model flexibility increases with the Number of learners setting.
All ensemble classifiers tend to be slow to fit because they often need many learners.
Classifier Type  Prediction Speed  Memory Usage  Interpretability  Ensemble Method  Model Flexibility 

Boosted Trees  Fast  Low  Hard  AdaBoost , with Decision
Tree learners  Medium to high — increases with Number of learners or Maximum number of splits setting.
Tip Boosted trees can usually do better than bagged, but might require parameter tuning and more learners

Bagged Trees  Medium  High  Hard  Random forestBag ,
with Decision Tree learners  High — increases with Number of learners setting.
Tip Try this classifier first.

Subspace
Discriminant  Medium  Low  Hard  Subspace , with
Discriminant learners  Medium — increases with Number of learners setting. Good for many predictors 
Subspace KNN  Medium  Medium  Hard  Subspace , with Nearest
Neighbor learners  Medium — increases with Number of learners setting. Good for many predictors 
RUSBoost Trees  Fast  Low  Hard  RUSBoost , with Decision
Tree learners  Medium — increases with Number of learners or Maximum number of splits setting. Good for skewed data (with many more observations of 1 class) 
GentleBoost or LogitBoost — not available in the
Model Type gallery. If you have 2 class data, select manually.  Fast  Low  Hard  GentleBoost or
LogitBoost , with
Decision Tree
learnersChoose Boosted
Trees and change to
GentleBoost method.  Medium — increases with Number of learners or Maximum number of splits setting. For binary classification only 
Bagged trees use Breiman's 'random forest'
algorithm. For
reference, see Breiman, L. Random Forests. Machine Learning 45, pp. 5–32,
2001.
Tips
Try bagged trees first. Boosted trees can usually do better but might require searching many parameter values, which is timeconsuming.
Try training each of the nonoptimizable ensemble classifier options in the Model Type gallery. Train them all to see which settings produce the best model with your data. Select the best model in the Models pane. To try to improve your model, try feature selection, PCA, and then (optionally) try changing some advanced options.
For boosting ensemble methods, you can get fine detail with either deeper trees or larger numbers of shallow trees. As with single tree classifiers, deep trees can cause overfitting. You need to experiment to choose the best tree depth for the trees in the ensemble, in order to tradeoff data fit with tree complexity. Use the Number of learners and Maximum number of splits settings.
For an example, see Train Ensemble Classifiers Using Classification Learner App.
Ensemble classifiers in Classification Learner use the fitcensemble
function. You can
set these options:
For help choosing Ensemble method and Learner type, see the Ensemble table. Try the presets first.
Maximum number of splits
For boosting ensemble methods, specify the maximum number of splits or branch points to control the depth of your tree learners. Many branches tend to overfit, and simpler trees can be more robust and easy to interpret. Experiment to choose the best tree depth for the trees in the ensemble.
Number of learners
Try changing the number of learners to see if you can improve the model. Many learners can produce high accuracy, but can be time consuming to fit. Start with a few dozen learners, and then inspect the performance. An ensemble with good predictive power can need a few hundred learners.
Learning rate
Specify the learning rate for shrinkage. If you set the learning rate to less than 1, the ensemble requires more learning iterations but often achieves better accuracy. 0.1 is a popular choice.
Subspace dimension
For subspace ensembles, specify the number of predictors to sample in each learner. The app chooses a random subset of the predictors for each learner. The subsets chosen by different learners are independent.
Alternatively, you can let the app choose some of these model options automatically by using hyperparameter optimization. See Hyperparameter Optimization in Classification Learner App.
Neural network models typically have good predictive accuracy and can be used for multiclass classification; however, they are not easy to interpret.
Model flexibility increases with the size and number of fully connected layers in the neural network.
Tip
In the Model Type gallery, click All Neural Networks to try each of the preset neural network options and see which settings produce the best model with your data. Select the best model in the Models pane, and try to improve that model by using feature selection and changing some advanced options.
Classifier Type  Prediction Speed  Memory Usage  Interpretability  Model Flexibility 

Narrow Neural
Network  Fast  Medium  Hard  Medium — increases with Size of each fully connected layer setting 
Medium Neural
Network  Fast  Medium  Hard  Medium — increases with Size of each fully connected layer setting 
Wide Neural Network  Fast  Medium  Hard  Medium — increases with Size of each fully connected layer setting 
Bilayered Neural
Network  Fast  Medium  Hard  High — increases with Size of each fully connected layer setting 
Trilayered Neural
Network  Fast  Medium  Hard  High — increases with Size of each fully connected layer setting 
Each model is a feedforward, fully connected neural network for classification. The first fully connected layer of the neural network has a connection from the network input (predictor data), and each subsequent layer has a connection from the previous layer. Each fully connected layer multiplies the input by a weight matrix and then adds a bias vector. An activation function follows each fully connected layer. The final fully connected layer and the subsequent softmax activation function produce the network's output, namely classification scores (posterior probabilities) and predicted labels. For more information, see Neural Network Structure.
For an example, see Train Neural Network Classifiers Using Classification Learner App.
Neural network classifiers in Classification Learner use the fitcnet
function. You can set these options:
Number of fully connected layers — Specify the number of fully connected layers in the neural network, excluding the final fully connected layer for classification. You can choose a maximum of three fully connected layers.
Size of each fully connected layer — Specify the size of each fully connected layer, excluding the final fully connected layer. If you choose to create a neural network with multiple fully connected layers, consider specifying layers with decreasing sizes.
Activation — Specify the activation function for
all fully connected layers, excluding the final fully connected layer.
The activation function for the last fully connected layer is always
softmax. Choose from the following activation functions:
ReLU
, Tanh
,
Sigmoid
, and None
.
Iteration limit — Specify the maximum number of training iterations.
Regularization term strength — Specify the ridge (L2) regularization penalty term.
Standardize data — Specify whether to standardize the numeric predictors. If predictors have widely different scales, standardizing can improve the fit. Standardizing the data is highly recommended.