Introduction to Supervised Machine Learning

Supervised machine learning is a subfield of machine learning (ML) that deals with building models from labeled data in order to predict the outcomes for unseen data.

A labeled data set consists of a set of examples that have been “tagged” or labeled with the correct output. In this regard, supervised learning can be compared to learning from a teacher that provides a set of examples with the correct answers.

For example, we can use supervised learning to build a spam filter by learning from a large amount of emails that have been tagged as spam or non-spam (ham). The resulting filter would be more effective and more robust than a spam filter that is built from manual definitions of if-then rules that check for different patterns in the email (rules that tend to become obsolete over time as new spamming strategies evolve).

Applications based on supervised machine learning are involved in almost every aspect of our lives, including:

• Fraud detection systems protect banks from malicious attackers.
• Medical diagnosis systems can diagnose skin cancer better than dermatologists.
• Speech recognition systems can translate voice to text more accurately and faster than humans.
• Advertising systems learn to match the right ads with the right context.

This article formally defines the supervised machine learning problem, describes the main steps involved in building a…