Machine Learning Anomaly Detectors

Creation and workflow for time series machine learning anomaly detectors

The machine learning anomaly detectors are based on unsupervised detection algorithms in Statistics and Machine Learning Toolbox™. These algorithms use different methods to identify outliers within a set of data.

Machine learning algorithms tend to be relatively fast, and are often good detectors to start with when you begin the process of finding the right detector for your data.

Apps

Time Series Anomaly Detector

Interactively create, train, test, and tune detectors for detecting anomalous behavior in time series (Since R2026a)

Functions

expand all

Create Machine Learning Detectors

`timeSeriesIforestAD`	Create a machine learning isolation forest anomaly detector model for time series data (Since R2026a)
`timeSeriesRrcforestAD`	Create a machine learning robust random cut forest anomaly detector model for time series data (Since R2026a)
`timeSeriesLofAD`	Create a machine learning local outlier factor model for anomaly detection in time series data (Since R2026a)
`timeSeriesOcsvmAD`	Create a machine learning one-class SVM anomaly detector model for time series data (Since R2026a)

Train and Test Sequence for Machine Learning Detectors

`train`	Train time series machine learning anomaly detector and obtain detection threshold (Since R2026a)
`detect`	Detect anomalies in time series using a trained time series machine learning detector model (Since R2026a)
`updateDetector`	Update settings of a trained time series machine learning anomaly detector and recompute detection threshold (Since R2026a)

Visualize and Evaluate Detector Performance

`plotHistogram`	Plot histogram of anomaly scores and detection threshold for trained machine learning anomaly detector (Since R2025a)
`plot`	Plot detected anomalies and anomaly scores generated from trained machine learning anomaly detectors (Since R2026a)
`timeSeriesAnomalyMetrics`	Compute specialized evaluation metrics for time series anomaly detection (Since R2026a)

Machine Learning Detector Model Objects

`TimeSeriesIForestDetector`	Detect subsequence anomalies in time series using an isolation forest algorithm (Since R2026a)
`TimeSeriesRRCForestDetector`	Detect subsequence anomalies in time series using a robust random cut forest algorithm (Since R2026a)
`TimeSeriesLOFDetector`	Detect subsequence anomalies in time series using a local outlier factor algorithm (Since R2026a)
`TimeSeriesOCSVMDetector`	Detect subsequence anomalies in time series using a one-class SVM detector (Since R2026a)

Topics

Detecting Anomalies in Time Series
Examine the general workflow for developing anomaly detectors that detect anomalous subsequences in time series.
Train and Test Isolation Forest Time Series Anomaly Detector
This example shows the development cycle for a Machine Learning anomaly detector.
Interpret Evaluation Metrics for Time Series Anomaly Detectors
Interpret evaluation metrics that are returned by the evaluationMetrics function and the app.