![](https://csdnimg.cn/release/download_crawler_static/11945302/bg11.jpg)
xvi Preface
packages, together with an accessible, intuitive, and highly efficient
base language, makes Julia a formidable platform for data science.
This book is not intended as an exhaustive introduction to data
science. In fact, this book is far from an exhaustive introduction to
data science. There are many very good books that one can consult
to learn about different aspects of data science (e.g., Bishop, 2006;
Hastie et al., 2009; Schutt, 2013; White, 2015; Efron and Hastie,
2016), but this book is primarily about Julia. Nevertheless, sev-
eral important topics in data science are covered. These include
data visualization, supervised learning, and unsupervised learn-
ing. When discussing supervised learning, we place some focus on
gradient boosting — a machine learning technique — because we
have found this approach very effective in applications. However,
for unsupervised learning, we take a more statistical approach and
place some focus on the use of probabilistic principal components
analyzers and a mixture thereof.
This monograph is laid out to progress naturally. In Chapter 1,
we discuss data science and provide some historical context. Julia
is also introduced as well as details of the packages and datasets
used herein. Chapters 2 and 3 cover the basics of the Julia language
was well as how to work with data in Julia. After that (Chapter 4),
a crucially important topic in data science is discussed: visualiza-
tion. The book continues with selected techniques in supervised
(Chapter 5) and unsupervised learning (Chapter 6), before con-
cluding with details of how to call R functions from within Julia
(Chapter 7). This last chapter also provides further examples of
mixture model-based clustering as well as an example that uses
random forests. Some appendices are included to provide readers
with some relevant mathematics, Julia performance tips and a list
of useful linear algebra functions in Julia.
There is a large volume of Julia code throughout this book,
which is intended to help the reader gain familiarity with the lan-
guage. We strongly encourage readers to run the code for them-
selves and play around with it. To make the code as easy as possible
to work with, we have interlaced it with comments. As readers be-
gin to get to grips with Julia, we encourage them to supplement
or replace our comments with their own. For the reader’s con-
venience, all of the code from this book is available on GitHub:
github.com/paTait/dswj.
We are most grateful to David Grubbs of the Taylor & Francis
Group for his support in this endeavour. His geniality and pro-
fessionalism are always very much appreciated. Special thanks to
E-Books & Papers for Statisticians