
Classification
Classification models are supervised learning methods and are aimed at predicting a categorical target. From a set of observations for which the class is known, a model that allows us to make predictions is generated.
The Statistics and Machine Learning Toolbox offers apps and functions that cover a variety of parametric and non-parametric classification algorithms, such as:
- Logistic regression
- Boosted and bagged decision trees, including AdaBoost, LogitBoost, GentleBoost, and RobustBoost
- Naive Bayes classification
- KNN classification
- Discriminant analysis (linear and quadratic)
- SVM (binary and multiclass classification)
The Classification Learner app is a very useful tool that executes more request activities such as interactively explore data, select features, specify cross-validation schemes, train models, and assess results. We can use it to perform common tasks such as:
- Importing data and specifying cross-validation schemes
- Exploring data and selecting features
- Training models using several classification algorithms
- Comparing and assessing models
- Sharing trained models for use in applications such as computer vision and signal processing
Using the Classification Learner app, we can choose between various algorithms to train and validate classification models. After the training, compare the models' validation errors and choose the best one on the basis of results.
The following figure shows the Classification Learner app:

Figure 1.18: The Classification Learner with a history list containing various classifier types