Mastering Machine Learning on AWS: Advanced machine learning in Python using SageMaker, Apache Spark, and TensorFlow

293 Pages·2019·19.447 MB·English

by  Dr. Saket S.R. Mengle, Maximo Gurmendez

Mastering Machine Learning on
AWS
Advanced machine learning in Python using SageMaker,
Apache Spark, and TensorFlow
Dr. Saket S.R. Mengle
Maximo Gurmendez
BIRMINGHAM - MUMBAI
Mastering Machine Learning on AWS
Copyright © 2019 Packt Publishing
All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form
or by any means, without the prior written permission of the publisher, except in the case of brief quotations
embedded in critical articles or reviews.
Every effort has been made in the preparation of this book to ensure the accuracy of the information presented.
However, the information contained in this book is sold without warranty, either express or implied. Neither the
authors, nor Packt Publishing or its dealers and distributors, will be held liable for any damages caused or alleged to
have been caused directly or indirectly by this book.
Packt Publishing has endeavored to provide trademark information about all of the companies and products
mentioned in this book by the appropriate use of capitals. However, Packt Publishing cannot guarantee the accuracy
of this information.
Commissioning Editor: Sunith Shetty
Acquisition Editor: Devika Battike
Content Development Editor: Nathanya Dias
Technical Editor: Utkarsha S. Kadam
Copy Editor: Safis Editing
Project Coordinator: Kirti Pisat
Proofreader: Safis Editing
Indexer: Priyanka Dhadke
Graphics: Jisha Chirayil
Production Coordinator: Shraddha Falebhai
First published: May 2019
Production reference: 1150519
Published by Packt Publishing Ltd.
Livery Place
35 Livery Street
Birmingham
B3 2PB, UK.
ISBN 978-1-78934-979-5
www.packtpub.com
I would like to dedicate this book in memory of my dad. Thanks for being there for me and
supporting my dreams.
– Dr. Saket S.R. Mengle
This book is dedicated to Mateo and Paulina, who are my constant source of inspiration, joy and
purpose.
– Maximo Gurmendez
mapt.io
Mapt is an online digital library that gives you full access to over 5,000 books and videos, as
well as industry leading tools to help you plan your personal development and advance
your career. For more information, please visit our website.
Why subscribe?
Spend less time learning and more time coding with practical eBooks and Videos
from over 4,000 industry professionals
Improve your learning with Skill Plans built especially for you
Get a free eBook or video every month
Mapt is fully searchable
Copy and paste, print, and bookmark content
Packt.com
Did you know that Packt offers eBook versions of every book published, with PDF and
ePub files available? You can upgrade to the eBook version at www.packt.com and as a print
book customer, you are entitled to a discount on the eBook copy. Get in touch with us at
[email protected] for more details.
At www.packt.com, you can also read a collection of free technical articles, sign up for a
range of free newsletters, and receive exclusive discounts and offers on Packt books and
eBooks.
Contributors
About the authors
Dr. Saket S.R. Mengle holds a PhD in text mining from Illinois Institute of Technology,
Chicago. He has worked in a variety of fields, including text classification, information
retrieval, large-scale machine learning, and linear optimization. He currently works as
senior principal data scientist at dataxu, where he is responsible for developing and
maintaining the algorithms that drive dataxu's real-time advertising platform.
I would like to thank my wife, Sharvari, who gives me strength and inspires me to be the
best version of myself every day. This book would have not been possible without her love
and support. I would also like to thank my parents, Subhash and Rashmi Mengle, who
taught me the value of hard work. I would like to express my appreciation to my advisor,
Dr. Nazli Goharian, and Dr. Ophir Frieder, who introduced me to the world of Machine
Learning.
Maximo Gurmendez holds a master's degree in computer science/AI from Northeastern
University, where he attended as a Fulbright Scholar. Since 2009, he has been working with
dataxu as data science engineering lead. He's also the founder of Montevideo Labs (a data
science and engineering consultancy). Additionally, Maximo is a computer science
professor at the University of Montevideo and is director of its data science for business
program.
I'd like to deeply thank my wife Maggie for her sustained support, encouragement and
patience, especially throughout the long work days and busy weekends that writing this
book implied. Additionally, I’d like to thank my mother, Margarita, who taught me the
importance of learning, caring and hard-work through her own example. Finally, I’d like to
express my gratitude to the dataxu team from whom I learned so much in the past ten
years.
About the reviewer
Chirag Nayyar helps organizations initiate their digital transformation using the public
cloud. He has been actively working on cloud platforms since 2013, providing consultancy
to many organizations, ranging from small and mid-size businesses to enterprises. He
holds a wide range of certifications from all major public cloud platforms. He also runs a
meet-up group and is a regular speaker at various cloud events. He has also reviewed
Hands-On Machine Learning on Google Cloud Platform and Google Cloud Platform Cookbook, by
Packt Publishing.
Packt is searching for authors like you
If you're interested in becoming an author for Packt, please visit authors.packtpub.com
and apply today. We have worked with thousands of developers and tech professionals,
just like you, to help them share their insight with the global tech community. You can
make a general application, apply for a specific hot topic that we are recruiting an author
for, or submit your own idea.
Table of Contents
Preface 1
Section 1: Machine Learning on AWS
Chapter 1: Getting Started with Machine Learning for AWS 9
How AWS empowers data scientists 9
Using AWS tools for machine learning 11
Identifying candidate problems that can be solved using machine
learning 12
Machine learning project life cycle 13
Data gathering 13
Evaluation metrics 14
Algorithm selection 14
Deploying models 15
Summary 15
Exercise 15
Section 2: Implementing Machine Learning
Algorithms at Scale on AWS
Chapter 2: Classifying Twitter Feeds with Naive Bayes 17
Classification algorithms 18
Feature types 19
Nominal features 19
Ordinal features 19
Continuous features 19
Naive Bayes classifier 20
Bayes' theorem 20
Posterior 21
Likelihood 21
Prior probability 21
Evidence 22
How the Naive Bayes algorithm works 22
Classifying text with language models 24
Collecting the tweets 24
Preparing the data 25
Building a Naive Bayes model through SageMaker notebooks 26
Naïve Bayes model on SageMaker notebooks using Apache Spark 33
Using SageMaker's BlazingText built-in ML service 36
Naive Bayes – pros and cons 40
Table of Contents
Summary 41
Exercises 42
Chapter 3: Predicting House Value with Regression Algorithms 43
Predicting the price of houses 43
Understanding linear regression 44
Linear least squares estimation 46
Maximum likelihood estimation 47
Gradient descent 47
Evaluating regression models 48
Mean absolute error 48
Mean squared error 49
Root mean squared error 49
R-squared 49
Implementing linear regression through scikit-learn 50
Implementing linear regression through Apache Spark 53
Implementing linear regression through SageMaker's linear Learner 55
Understanding logistic regression 59
Logistic regression in Spark 59
Pros and cons of linear models 60
Summary 60
Chapter 4: Predicting User Behavior with Tree-Based Methods 61
Understanding decision trees 61
Recursive splitting 63
Types of decision trees 63
Cost functions 64
Gini Impurity 64
Information gain 66
Criteria to stop splitting trees 67
Understanding random forest algorithms 68
Understanding gradient boosting algorithms 69
Predicting clicks on log streams 69
Introduction to Elastic MapReduce (EMR) 70
Training with Apache Spark on EMR 73
Getting the data 74
Preparing the data 74
Categorical encoding 78
One-hot encoding 78
Training a model 81
Evaluating our model 82
Area Under ROC Curve 83
Area under the precision-recall curve 84
Training tree ensembles on EMR 86
Training gradient-boosted trees with the SageMaker services 87
Preparing the data 87
Training with SageMaker XGBoost 88
[ ii ]
Table of Contents
Applying and evaluating the model 90
Summary 93
Exercises 94
Chapter 5: Customer Segmentation Using Clustering Algorithms 95
Understanding How Clustering Algorithms Work 95
k-means clustering 97
Euclidean distance 99
Manhattan distance 99
Hierarchical clustering 99
Agglomerative clustering 99
Divisive clustering 101
Clustering with Apache Spark on EMR 102
Clustering with Spark and SageMaker on EMR 111
Understanding the purpose of the IAM role 114
Summary 118
Exercises 118
Chapter 6: Analyzing Visitor Patterns to Make Recommendations 119
Making theme park attraction recommendations through Flickr data 119
Collaborative filtering 120
Memory-based approach 120
Model-based approach 121
Matrix factorization 121
Stochastic gradient descent 123
Alternating Least Squares 123
Finding recommendations through Apache Spark's ALS 124
Data gathering and exploration 124
Training the model 127
Getting recommendations 128
Recommending attractions through SageMaker Factorization
Machines 131
Preparing the dataset for learning 131
Training the model 135
Getting recommendations 137
Summary 139
Exercises 140
Section 3: Deep Learning
Chapter 7: Implementing Deep Learning Algorithms 142
Understanding deep learning 142
Applications of deep learning 143
Self-driving cars 144
Learning to play video games using a deep learning algorithm 145
Understanding deep learning algorithms 145
[ iii ]