Summary

In this chapter, we saw how ML.NET gives us different complexity options depending on our needs.

In its simplest form, you can run an AutoML experiment in a single line of code and accept the default pipeline and model trainer options. You can also get more granular control by building your own pipeline but still rely on ML.NET AutoML to tune hyperparameters for you, or select the appropriate model trainer. Finally, you can get complete control by working directly with ML.NET model trainers and not using AutoML at all.

We saw how different model trainers have different hyperparameters, and each hyperparameter has different ranges of values that make sense for it. These combinations of hyperparameter values constitute the search space that you must navigate to find optimal values for your hyperparameters, through hyperparameter tuning.

We discussed a wide range of machine learning algorithms for regression and classification. We also saw how some linear models need...