Setareh Rafatirad - Machine Learning for Computer Scientists and Data Analysts: From an Applied Perspective

Chapter introduces the concept of applied machine learning. The metrics used for evaluating the machine learning performance, data pre-processing, and techniques to visualize and analyze the outputs (classification or regression or other applications) are discussed.

Chapter presents a brief review of the probability theory and linear algebra that are essential for a better understanding of the ML techniques discussed in the later parts of the book.

Chapter introduces the machine learning techniques. Supervised learning is primarily discussed in this chapter. Multiple supervised learning techniques, learning techniques, and applications along with pros and cons for each of the techniques are discussed. A qualitative comparison of different supervised learning techniques is presented along with their suitability to different kinds of applications.

Unsupervised learning is introduced in Chap.. The differences compared to the supervised learning and application scenarios are discussed first. Different supervised learning for different applications including classification and feature selection is discussed along with examples in this chapter.

Reinforcement learning is a human learning-inspired technique, which can be laid between supervised and unsupervised learning techniques in the spectrum. Chapter discusses the basics of reinforcement learning along with its variants together with a comparison among different techniques.

The majority of the supervised learning techniques and their learning mechanisms discussed in the first part of this book focus on offline or batch learning. However, the learning in real-world applications needs to happen in an online manner. As such, Chap. introduces the online learning technique and different variants of online learning techniques.

With a diverse spectrum of Web applications demonstrating the importance of learning from user behavior, recommender systems are widely used by the bulk of social media companies. Chapter of this book discusses approaches for recommender learning.

Chapter offers approaches for graph learning. Graphs are used to depict things and their connections in a variety of real-world applications, including social networking, transportation, and disease spreading. Methods for learning graphs and the relationships between nodes are discussed.

In addition to advancements in machine learning algorithms, researchers have also focused on exploiting the vulnerabilities in machine learning techniques. Chapter introduces adversarial machine learning techniques that discuss techniques to inject the adversarial perturbations into the input samples to mislead the machine learning algorithms. In addition, the techniques to harden the machine learning techniques against these adversarial perturbations are discussed.

The application of machine learning techniques for health monitoring is one of the critical real-world applications, especially with the introduction of wearable devices including fitness trackers. Chapters focus on the application of machine learning techniques for health applications, particularly in the context of wearable devices.