Other > E-books
Textbook in PDF format
Machine learning (ML) has reached its maturity, due to the advancements in both computing power and new theoretical and algorithmic development in the recent decades. The technology has found many new applications in a wide spectrum of areas in both industry and our daily life, thereby generating passion among college students as well as practicing professionals with different backgrounds to gain the knowledge and master the skill in it, as a fascinating intersection of applied mathematics, computer science, and engineering.
This book attempts to bridge the gap between the theoretical foundation of the ML methods and their algorithmic and code level implementation, so that readers seriously interested in the subject can learn to understand both why and how the ML algorithms work.
The book is organized in six parts. Part I is mostly dedicated to optimization, covering constrained as well as unconstrained methods. Part II covers the most important regression algorithms widely used in ML, including learning and nonlinear regression, logistic and softmax regression, and Gaussian process regression. Part III is about feature selection and dimensionality reduction, as the preprocessing stage of the main ML operations. Part IV covers a set of core classification methods widely used in ML, including both supervised methods and unsupervised clustering. Part V introduces a set of important algorithms based on artificial neural networks inspired by the biological neural network in the brain. The last part provides an introduction to the reinforcement learning.
This book is motivated by the author's teaching experience over several years in China University of Petroleum (CUP), and it grew out of the lecture notes developed for the course. It is written for undergraduate or graduate students in computer science, engineering, or any other relevant majors. It can also be used as a reference for practicing professionals of different background
