smarthi / anomaly_detection

Time series anomaly detection in Mahout

Introduction

This repository contains a new class for time series anomaly detection in Mahout and a corresponding example based on Ted Dunning's previous work on EKG data.

You can find the new class under src/main/java/org/apache/mahout/anomalydetection/TimeSeriesAnomalyDetection.java.

The TimeSeriesAnomalyDetection class embeds the t-digest algorithm in order to spot anomalies and guides the user through the process of anomaly detection.

The EKAnomalyDetection class implements a time series anomaly detection scenario by applying the newly introduced TimeSeriesAnomalyDetection class.

The example is provided under src/main/java/org/apache/mahout/anomalydetection/EKGAnomalyDetection.java.

How to run the example

In order to run the example:

1. Assure maven is installed in your system (https://maven.apache.org/)

2. Execute: mvn clean install

3. Execute the following command:

   mvn -q exec:java -Dexec.mainClass=org.apache.mahout.anomalydetection.EKGAnomalyDetection

In order to test it run: mvn test

References

For further information:

Anomaly detection

• Practical Machine Learning: A New Look At Anomaly Detection by Ted Dunning and Ellen Friedman
• A talk about anomaly detection
• Related to this example on anomaly detection on EKG data

t-digest algorithm

• The original implementation of and documentation for t-digest