Machine Learning and its Applications

205 Pages·2019·19.086 MB·English

by  Peter Wlodarczak

Machine Learning and
its Applications
Peter Wlodarczak
University of Southern Queensland
Toowoomba, Queensland, Australia
p,
p,
A SCIENCE PUBLISHERS BOOK
A SCIENCE PUBLISHERS BOOK
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2020 by Taylor & Francis Group, LLC
CRC Press is an imprint of Taylor & Francis Group, an Informa business
No claim to original U.S. Government works
Printed on acid-free paper
Version Date: 20190821
International Standard Book Number-13: 978-1-138-32822-8 (Hardback)
Th is book contains information obtained from authentic and highly regarded sources. Reasonable eff orts have been
made to publish reliable data and information, but the author and publisher cannot assume responsibility for the
validity of all materials or the consequences of their use. Th e authors and publishers have attempted to trace the
copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to
publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let
us know so we may rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted,
or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, includ-
ing photocopying, microfi lming, and recording, or in any information storage or retrieval system, without written
permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www.copyright.com
(http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers,
MA 01923, 978-750-8400. CCC is a not-for-profi t organization that provides licenses and registration for a variety
of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment
has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for
identifi cation and explanation without intent to infringe.
Library of Congress Cataloging-in-Publication Data
Names: Wlodarczak, Peter, author.
Title: Machine learning and its applications / Peter Wlodarczak, University
of Southern Queensland, Toowoomba, Queensland, Australia.
Description: First edition. | Boca Raton, FL : CRC Press/Taylor & Francis
Group, [2020] | Includes bibliographical references and index. |
Summary: “This book describes Machine Learning techniques and algorithms
that have been used in recent real-world application. It provides an
introduction to Machine Learning, describes the most widely used
techniques and methods. It also covers Deep Learning and related areas
such as function approximation or. The book gives real world examples
where Machine Learning techniques are applied and describes the basic
math and the commonly used learning techniques”-- Provided by publisher.
Identifiers: LCCN 2019033419 | ISBN 9781138328228 (hardcover ; acid-free
paper)
Subjects: LCSH: Machine learning.
Classification: LCC Q325.5 .W635 2020 | DDC 006.3/1--dc23
LC record available at https://lccn.loc.gov/2019033419
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
For my parents

Preface
Artificial intelligence or AI, and in particular, machine learning has
received a lot of attention in media due to the rapid advances we
have seen in this area over the past few years. Progress in artificial
intelligence has been slow in the 1990’s and a certain disillusionment
was noticeable in research since artificial intelligence did not seem
to live up to the expectation. Since the beginning of the new century,
machine learning has seen rapid progress due to new approaches, in
particular, deep learning. Deep learning goes further back, however,
due to recent advances in other areas, in particular, the development
of graphics processing units that provide the necessary computing
power, and the Internet, that made an unprecedented amount of data
publicly available, machine learning has gained a lot of momentum and
we have seen an incredible amount of new applications in many areas
of research and in practice. Machine learning is used for spam filtering,
for medical image analysis, for voice commands, for autonomously
driving cars and to analyze sensor data for the Internet of Things
just to name a few. Every high-end smart phone is now AI enabled.
Deep learning is used to predict earthquakes, for automatic language
translation, for automatic coloring of black and white movies and
for forecasting in finance. The Mars rover Curiosity utilizes artificial
intelligence to autonomously select inspection-worthy soil and rock
samples with high accuracy. The list goes on.
Recently, machine learning also attracted a lot of attention due to
visions of jobs being lost to machines, dangers of autonomous systems,
vi < Machine Learning and its Applications
such as robots who could turn evil, loss of control over existential
infrastructure or even a war starting due to artificial intelligence.
However, it should be noted that artificial intelligence, as the name
suggests, is artificial and has nothing to do with human intelligence. A
machine may be capable of recognizing whether an animal in a picture
is a lion or a tiger, but it is not able to understand the concept behind
what it recognized, i.e., it does not understand the concept of a living
being. A deep learner is, in essence, a huge mathematical formula. It
has nothing to do with how the human brain works, where biochemical
processes are executed. An artificial neural network is inspired by
nature. It has a grain of how we think the brain works and a lot of math
thrown at it. Visions of machines replacing doctors in diagnosing a
patient or lawyers in sentencing a perpetrator are probably premature.
It is more likely that we will see artificial intelligence supporting a
doctor or a lawyer in his daily work instead of replacing her or him in
the near future.
However, we will likely see many more applications of artificial
intelligence in areas that we have not thought of yet. Artificial
intelligence and machine learning pose many interesting problems in
practice and in research. The fact that we try to imitate nature gives
us a better understanding of what problems nature had to solve during
evolution. It also helps us to better understand how it solved them.
However, how human learning works is still largely unknown. Before
we have a better understanding of how the brain works, it is unlikely
that we will be able to replace it, no matter how much math we apply.
When I started to work with machine learning there were excellent
books explaining the math used in machine learning algorithms. There
were also excellent books about data mining and knowledge discovery.
Some of them are listed in the reference section. They can be used for
studying and for reference. However, there were no books explaining
the basic concepts behind machine learning and how they are applied
in practice. It is important to understand the math and the algorithms
used in machine learning. However, it is not necessary to know all
the details about the algorithms to be able to effectively use them for
real world problems. Machine learning has evolved over the past years
to the point where a lot of the complexity is hidden in frameworks
such as TensorFlow or Apache Mahout. A data scientist can use these
Preface < vii
frameworks without knowing the implementation details. Many popular
programming languages have libraries with implementations that are
ready to be used in programs. R, Python, Java or Scala have many
different implementations of machine learning methods and the data
scientist can continue using his preferred programming language by
including the libraries that fit the purpose best. There are also tools, such
as RapidMiner, Weka or Knime to name a few, where no programming
skills are required and machine learning workflows can be created
graphically. Obviously, next to the open source implementations, there
are also commercial tools offered by different vendors to choose from.
This book aims to explain the basic concepts behind machine learning
and the machine learning methods. It is the work of several years of
experience in applying these methods in practice in various projects. It
tries to give a concise description of the algorithms and the math behind
them to the level where it helps explain the inner workings. However,
it is not intended to give an exhaustive mathematical description with
all the derivations. Also, machine learning goes back to the 1940’s
and has evolved since then and many different techniques have been
developed, too many to describe in one single book. Also, for virtually
every machine learning method, variations have been proposed that
might be more suitable for certain problems than others. Nevertheless,
some of the basic concepts apply to many of these methods and
understanding them makes it easy to familiarize oneself with new
methods that have not been used before. This book is written in such
a way that each chapter can be read individually with the caveat that
some redundancy exists in the chapters. This book will hopefully make
it easier for the reader to get started with machine learning and support
them in the fascinating journey through the world of data science and
machine learning.

Contents
Preface v
List of Figures xiii
List of Tables xv
SECTION I: INTRODUCTION
1. Introduction 3
1.1 Data mining 5
1.2 Data mining steps 6
1.3 Data collection 7
1.4 Data pre-processing 8
1.5 Data analysis 10
1.5.1 Supervised learning 10
1.5.2 Unsupervised learning 11
1.5.3 Semi-supervised learning 12
1.5.4 Machine learning and statistics 13
1.6 Data post-processing 15
2. Machine Learning Basics 17
2.1 Supervised learning 19
2.1.1 Perceptron 21
2.2 Unsupervised learning 25
2.2.1 k-means clustering 27