The goal of all machine learning (ML) activity is to turn raw data into some prediction or classification or insight. Raw data appears on the left or at the beginning of this pipeline, and on the right or at the end comes the insight/prediction/classification and so on. Although each machine learning task will require a different pipeline, the basic structure or the building blocks remain the same. ML.NET offers several types/interfaces to make the creation of this pipeline easier. A broad understanding of these concepts will help you understand how ML.NET works under the hood.