Steema Numerics and Machine Learning tools

What’s Steema Numerics?

It is a .NET 5.0 software library written in the C# programming language that allows us, through machine learning algorithms (ML.Net), to detect anomalies or make predictions in order to anticipate possible unforeseen events through data analytics.

Steema Numerics includes three assemblies that can be used with the TeeChart NET Component Library or independently of it:

Base library designed to facilitate the integration of Machine Learning into the analysis of your data.

To be able to use the library inside Windows.Forms applications.

To take advantage of Steema Numerics with TeeChart to create your visualisations.

Who is it made for?

For developers and software creators that are involved in data analytics, machine learning, Industry 4.0, among others, who want to integrate tools and new options for detecting problems in the data that their applications are manipulating.

Steema Numerics has been designed so that our current TeeChart.NET customers can integrate and use the tool together to show data results in an easy way.

It can also be used with other Chart components, making use only of the data work library with ML.Net and then displaying the results in the desired Chart component.

How can it help you?

Steema Numerics is designed to access and use the ML.Net machine learning algorithms (ML.Net) for anomaly detection and prediction. It provides an easy way to plug-in to the datasource and visualise results.

Through a treatment of small or large amounts of data, we can obtain results with information on the anomalies found or the predictions, so being able to generate reports.

What value does it add to ML.NET?

We’ve done the work in configuring ML.NET’s algorithms to produce reliable results using Steema Numeric’s default settings. Although it takes time to understand and correctly use both ML.NET’s anomaly and prediction algorithms, this is especially true of the anomaly algorithm, a novel calculation produced by a team within Microsoft with whom we had contact to fine-tune its parameters. 

What predictive options allows us to configure?

Best Guess :

  • Overlay
  • Projected
  • None

Number of values to Predict
SSaForecastingEstimator parameters :

  • WindowSize
  • TrainSize
  • Series Length
  • ConfidentLevel

Resources :

Steema Numerics is available through nuget packages, you can find them here :

Steema Numerics package :

Steema Numerics for Windows.Forms package :

Steema Numerics for Windows.Forms and TeeChart for NET package :

Examples can be found at the Steema github account :