Introducing ML.NET and AutoML

There are many ways of training machine learning models, including Python, R, and cloud tools such as Azure Machine Learning Studio or Amazon SageMaker, but the best way of using .NET to train or deploy a machine learning model is with ML.NET.

ML.NET is an open-source library designed to train, save, and load machine learning models to accomplish a variety of tasks. We actually worked with ML.NET earlier in this book through the DataFrame class, which is maintained in the same GitHub repository as ML.NET (https://github.com/dotnet/machinelearning).

You can use ML.NET to train models, evaluate their performance, and use them to make predictions. Because ML.NET is a .NET library, it can be integrated directly into other .NET applications, including web applications, Azure functions, and cross-platform desktop and mobile applications.

In fact, because ML.NET supports Open Neural Network Exchange (ONNX) file formats, you can train a model in Python...