Machine Learning with scikit-learn Quick Start Guide

Splitting the data into training and test sets

The idea of training and test sets is fundamental to every machine learning problem. When thinking about this concept, it is easy to understand why the concept was introduced. Think of machine learning as the direct equivalent to the process of human learning; when learning a concept in mathematics, we first learn how to solve a set of problems with solutions attached to them so that we can understand the exact methodology involved in solving these problems. We then take a test at school or university and attempt to solve problems that we have not encountered or seen before in order to test our understanding.

The training set is a part of the dataset that a machine learning algorithm uses to ...