In a general sense, we may argue that a machine learns anytime it changes its internal structure, code, or data (in response to its inputs or external information) in a way that enhances its predicted future performance. Several of these modifications, such adding a new record to a data, are more appropriately dealt with in other fields and are not always more understood because they are referred to as “learning.” This book aims to provide the reader an overview of the wide variety of applications that may be reduced to a machine learning problem, and to do so in a way that provides some degree of order to an otherwise chaotic zoo of problems. After that, the book delves into some elementary tools from probability and statistics, which serve as a basic vocabulary for framing the sorts of machine learning problems that may be solved. The book concludes with a discussion of several basic but useful algorithms for tackling the difficult problem of classifying data. The next section will focus on more advanced tools and broad problems. This book’s primary objective is to present readers to various important machine learning algorithms. The authors deliberated over which algorithms to present, ultimately settling on those that are both widely used in practise and representative of a broad range of learning approaches. Since the world has gotten more “digitised” in recent years, and the amount of data accessible for learning has similarly increased, we also pay special attention to algorithms suitable for large scale learning (“Big Data”). Thus, in many applications, data is not a limiting factor while computation time is. As a result, we make a point of precisely defining the quantity of data and duration of computational time required to master a certain notion.