Anomaly Detection

Introduction

A common application of the Steema Numerics Anomaly detection is to time based (realtime) data.

Once a connection has been made to the input data, Steema.Numerics sets the best WindowSize for it, determining whether the data spans and is in the order of years, months, weeks, days, hours or minutes or less.

Once determined, Steema.Numerics calls ML.Net's DetectAnomaly functions, detectanomalybysrcnn, returning the result to the client process.

Using Steema.Numeric's Anomaly detection for your TeeChart Series

Add references to Steema.Numerics

Include Steema.Numerics dependencies in your application project: - Steema.Numerics - Steema.Numerics.ChartHelper - Steema.Numerics.Windows

Steema.Numerics may be referenced via Nuget.

Call the Anomaly algorithm

Steema.Numerics will return a an integer list of the indices of every series point, from the input series, that is considered an anomaly according the algorithm.

1. Add a Series to you chart to 'house/show' the Anomalies. This would typically be a Point Series.
2. Create a new integer array for the anomalies indices
3. Call the TeeChartHelper static method to generate the marker list.

   int[] anomalies = null;
   anomalies = TeeChartHelper.Anomalies_FromSeries(_series1);
   //where _series1 is the data series you wish to analyse.
   

4. Call the TeeChartHelper static method to add the anomalies to the chart.

   TeeChartHelper.Add_Anomalies(_series1, _series2, anomalies);
   //where _series2 is the new, empty, point series you wish to use to show the anomalies.