Incremental Learning

Fit classification model to streaming data and track its performance

Incremental learning, or online learning, involves processing incoming data from a data stream, possibly given little to no knowledge of the distribution of the predictor variables, aspects of the objective function, and whether the observations are labeled. Incremental learning problems contrast with traditional machine learning methods, in which enough labeled data is available to fit to a model, perform cross-validation to tune hyperparameters, and infer the predictor distribution characteristics.

Incremental learning requires a configured incremental model. You can create and configure an incremental model directly by calling an object, for example incrementalClassificationLinear, or you can convert a supported traditionally trained model to an incremental learner by using incrementalLearner. After configuring a model and setting up a data stream, you can fit the incremental model to the incoming chunks of data, track the predictive performance of the model, or perform both actions simultaneously.

For more details, see Incremental Learning Overview.

You can also incrementally monitor for drift in concept data, such as classification error. First you need to configure the drift detector using incrementalConceptDriftDetector. After setting up a data stream, you can update the drift detector and check for any drift using detectdrift. For more information, see the reference pages.

Blocks

IncrementalClassificationLinear Predict	Classify observations using incremental linear classification model (Since R2023b)
IncrementalClassificationLinear Fit	Fit incremental linear binary classification model (Since R2023b)
IncrementalClassificationECOC Predict	Classify observations using incremental ECOC classification model (Since R2024a)
IncrementalClassificationECOC Fit	Fit incremental ECOC classification model (Since R2024a)
IncrementalClassificationKernel Fit	Fit incremental kernel classification model (Since R2024b)
IncrementalClassificationKernel Predict	Classify observations using incremental kernel classification model (Since R2024b)
Update Metrics	Update performance metrics in incremental learning model given new data (Since R2023b)
Detect Drift	Update drift detector states and drift status with new data (Since R2024b)

Functions

expand all

Drift-Aware Learning

Create Incremental Drift-Aware Model

incrementalDriftAwareLearner Construct drift-aware model for incremental learning (Since R2022b)

Incrementally Fit and Track Performance

`fit`	Train drift-aware learner for incremental learning with new data (Since R2022b)
`updateMetrics`	Update performance metrics in incremental drift-aware learning model given new data (Since R2022b)
`updateMetricsAndFit`	Update performance metrics in incremental drift-aware learning model given new data and train model (Since R2022b)

Other Model Operations

`loss`	Regression or classification error of incremental drift-aware learner (Since R2022b)
`perObservationLoss`	Per observation regression or classification error of incremental drift-aware learner (Since R2022b)
`predict`	Predict responses for new observations from incremental drift-aware learning model (Since R2022b)
`reset`	Reset incremental drift-aware learner (Since R2022b)

Kernel Binary Classification

Create Incremental Model

`incrementalClassificationKernel`	Binary classification kernel model for incremental learning (Since R2022a)
`incrementalLearner`	Convert kernel model for binary classification to incremental learner (Since R2022a)

Incrementally Fit and Track Performance

`fit`	Train kernel model for incremental learning (Since R2022a)
`updateMetrics`	Update performance metrics in kernel incremental learning model given new data (Since R2022a)
`updateMetricsAndFit`	Update performance metrics in kernel incremental learning model given new data and train model (Since R2022a)

Other Model Operations

`predict`	Predict responses for new observations from kernel incremental learning model (Since R2022a)
`loss`	Loss of kernel incremental learning model on batch of data (Since R2022a)
`perObservationLoss`	Per observation classification error of model for incremental learning (Since R2022a)
`reset`	Reset incremental classification model (Since R2022a)

Linear Binary Classification

Create Incremental Model

`incrementalClassificationLinear`	Binary classification linear model for incremental learning (Since R2020b)
`incrementalLearner`	Convert binary classification support vector machine (SVM) model to incremental learner (Since R2020b)
`incrementalLearner`	Convert linear model for binary classification to incremental learner (Since R2020b)

Incrementally Fit and Track Performance

`fit`	Train linear model for incremental learning (Since R2020b)
`updateMetrics`	Update performance metrics in linear incremental learning model given new data (Since R2020b)
`updateMetricsAndFit`	Update performance metrics in linear incremental learning model given new data and train model (Since R2020b)

Other Model Operations

`predict`	Predict responses for new observations from linear incremental learning model (Since R2020b)
`loss`	Loss of linear incremental learning model on batch of data (Since R2020b)
`perObservationLoss`	Per observation classification error of model for incremental learning (Since R2022a)
`reset`	Reset incremental classification model (Since R2022a)

Multiclass ECOC Classification

Create Incremental Model

`incrementalClassificationECOC`	Multiclass classification model using binary learners for incremental learning (Since R2022a)
`incrementalLearner`	Convert multiclass error-correcting output codes (ECOC) model to incremental learner (Since R2022a)

Incrementally Fit and Track Performance

`fit`	Train ECOC classification model for incremental learning (Since R2022a)
`updateMetrics`	Update performance metrics in ECOC incremental learning classification model given new data (Since R2022a)
`updateMetricsAndFit`	Update performance metrics in ECOC incremental learning classification model given new data and train model (Since R2022a)

Other Model Operations

`predict`	Predict responses for new observations from ECOC incremental learning classification model (Since R2022a)
`loss`	Loss of ECOC incremental learning classification model on batch of data (Since R2022a)
`perObservationLoss`	Per observation classification error of model for incremental learning (Since R2022a)
`reset`	Reset incremental classification model (Since R2022a)

Naive Bayes Classification

Create Incremental Model

`incrementalClassificationNaiveBayes`	Naive Bayes classification model for incremental learning (Since R2021a)
`incrementalLearner`	Convert naive Bayes classification model to incremental learner (Since R2021a)

Incrementally Fit and Track Performance

`fit`	Train naive Bayes classification model for incremental learning (Since R2021a)
`updateMetrics`	Update performance metrics in naive Bayes incremental learning classification model given new data (Since R2021a)
`updateMetricsAndFit`	Update performance metrics in naive Bayes incremental learning classification model given new data and train model (Since R2021a)

Other Model Operations

`predict`	Predict responses for new observations from naive Bayes incremental learning classification model (Since R2021a)
`loss`	Loss of naive Bayes incremental learning classification model on batch of data (Since R2021a)
`logp`	Log unconditional probability density of naive Bayes classification model for incremental learning (Since R2021a)
`perObservationLoss`	Per observation classification error of model for incremental learning (Since R2022a)
`reset`	Reset incremental classification model (Since R2022a)

Concept Drift Detection

Create Concept Drift Detector

incrementalConceptDriftDetector Instantiate incremental concept drift detector (Since R2022a)

Detect Drift and Reset Model

`detectdrift`	Update drift detector states and drift status with new data (Since R2022a)
`reset`	Reset incremental concept drift detector (Since R2022a)

Objects

expand all

Incremental Concept Drift Detectors

`DriftDetectionMethod`	Incremental drift detector that utilizes Drift Detection Method (DDM) (Since R2022a)
`HoeffdingDriftDetectionMethod`	Incremental concept drift detector that utilizes Hoeffding's Bounds Drift Detection Method (HDDM) (Since R2022a)

Topics

Incremental Learning Overview
Discover fundamental concepts about incremental learning, including incremental learning objects, functions, and workflows.
Configure Incremental Learning Model
Prepare an incremental learning model for incremental performance evaluation and training on a data stream.
Implement Incremental Learning for Classification Using Succinct Workflow
Use the succinct workflow to implement incremental learning for binary classification with prequential evaluation.
Implement Incremental Learning for Classification Using Flexible Workflow
Use a flexible workflow to implement incremental learning for binary classification with prequential evaluation.
Initialize Incremental Learning Model from Logistic Regression Model Trained in Classification Learner
Train a logistic regression model using the Classification Learner app, and then initialize an incremental model for binary classification using the estimated coefficients.
Perform Conditional Training During Incremental Learning
Use the flexible workflow to implement conditional training during incremental learning with a naive Bayes multiclass classification model.
Perform Text Classification Incrementally
This example shows how to incrementally train a model to classify documents based on word frequencies in the documents; a bag-of-words model.
Incremental Learning with Naive Bayes and Heterogeneous Data
This example shows how to prepare heterogeneous predictor data, containing real-valued and categorical measurements, for incremental learning using a naive Bayes classifier.
Monitor Equipment State of Health Using Drift-Aware Learning
This example shows how to automate the process of monitoring the state of health for a cooling system using an incremental drift-aware learning model and Streaming Data Framework for MATLAB® Production Server™.

How useful was this information?

Unrated 1 star 2 stars 3 stars 4 stars 5 stars