Machine Learning Overview

Abstract

This chapter gives some background about machine learning as a whole and some of the techniques that touch all parts of it. "Machine learning" has become a catchall term that covers a lot of different areas, ranging from classification to clustering. Machine learning was partly born out of the initial failures of the artificial intelligence (AI) movement. Most machine learning algorithms are designed to handle pretty much any tabular dataset, but they ONLY handle tabular data. Tabular data lends itself to all kinds of mathematical analysis, since the rows of a table with n rows and d columns can be viewed as locations in d-dimensional space. This is why machine learning is easily the most mathematically sophisticated thing a data scientist is likely to do. Machine learning is much more probabilistic in the way it makes models and inferences. There are two main types of machine learning, called supervised and unsupervised.