Python Machine Learning Blueprints

371 Pages·2019·36.231 MB·English

by  Michael Roman, Alexander Combs;

Python Machine Learning
Blueprints
Second Edition
Put your machine learning concepts to the test by developing
real-world smart projects
Alexander Combs
Michael Roman
BIRMINGHAM - MUMBAI
Python Machine Learning Blueprints
Second Edition
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: Varsha Shetty
Content Development Editor: Snehal Kolte
Technical Editor: Naveen Sharma
Copy Editor: Safis Editing
Project Coordinator: Manthan Patel
Proofreader: Safis Editing
Indexer: Mariammal Chettiyar
Graphics: Jisha Chirayil
Production Coordinator: Arvindkumar Gupta
First published: July 2016
Second edition: January 2019
Production reference: 1310119
Published by Packt Publishing Ltd.
Livery Place
35 Livery Street
Birmingham
B3 2PB, UK.
ISBN 978-1-78899-417-0
www.packtpub.com
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
Alexander Combs is an experienced data scientist, strategist, and developer with
a background in financial data extraction, natural language processing and generation,
and quantitative and statistical modeling. He currently lives and works in New York City.
Writing a book is truly a massive undertaking that would not be possible without the
support of others. I would like to thank my family for their love and encouragement and
Jocelyn for her patience and understanding. I owe all of you tremendously.
Michael Roman is a data scientist at The Atlantic, where he designs, tests, analyzes, and
productionizes machine learning models to address a range of business topics. Prior to this
he was an associate instructor at a full-time data science immersive program in New York
City. His interests include computer vision, propensity modeling, natural language
processing, and entrepreneurship.
About the reviewer
Saurabh Chhajed is a machine learning and big data engineer with 9 years of professional
experience in the enterprise application development life cycle using the latest frameworks,
tools, and design patterns. He has experience of designing and implementing some of the
most widely used and scalable customer-facing recommendation systems with extensive
usage of the big data ecosystem— the batch, real-time, and machine learning pipeline. He
has also worked for some of the largest investment banks, credit card companies, and
manufacturing companies around the world, implementing a range of robust and scalable
product suites.
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
Chapter 1: The Python Machine Learning Ecosystem 6
Data science/machine learning workflow 7
Acquisition 8
Inspection 8
Preparation 9
Modeling 9
Evaluation 9
Deployment 10
Python libraries and functions for each stage of the data science
workflow 10
Acquisition 10
Inspection 11
The Jupyter Notebook 11
Pandas 13
Visualization 21
The matplotlib library 23
The seaborn library 29
Preparation 32
map 33
apply 34
applymap 35
groupby 36
Modeling and evaluation 38
Statsmodels 38
Scikit-learn 42
Deployment 47
Setting up your machine learning environment 48
Summary 48
Chapter 2: Build an App to Find Underpriced Apartments 49
Sourcing apartment listing data 50
Pulling down listing data 50
Pulling out the individual data points 57
Parsing data 61
Inspecting and preparing the data 64
Sneak-peek at the data types 66
Visualizing our data 69
Visualizing the data 77
Modeling the data 79
Table of Contents
Forecasting 82
Extending the model 86
Summary 86
Chapter 3: Build an App to Find Cheap Airfares 87
Sourcing airfare pricing data 88
Retrieving fare data with advanced web scraping 89
Creating a link 91
Parsing the DOM to extract pricing data 93
Parsing 96
Identifying outlier fares with anomaly detection techniques 106
Sending real-time alerts using IFTTT 113
Putting it all together 118
Summary 121
Chapter 4: Forecast the IPO Market Using Logistic Regression 122
The IPO market 123
What is an IPO? 123
Recent IPO market performance 124
Working on the DataFrame 126
Analyzing the data 131
Summarizing the performance of the stocks 132
Baseline IPO strategy 136
Data cleansing and feature engineering 139
Adding features to influence the performance of an IPO 139
Binary classification with logistic regression 141
Creating the target for our model 143
Dummy coding 143
Examining the model performance 145
Generating the importance of a feature from our model 149
Random forest classifier method 151
Summary 153
Chapter 5: Create a Custom Newsfeed 154
Creating a supervised training set with Pocket 155
Installing the Pocket Chrome Extension 155
Using the Pocket API to retrieve stories 157
Using the Embedly API to download story bodies 164
Basics of Natural Language Processing 166
Support Vector Machines 169
IFTTT integration with feeds, Google Sheets, and email 172
Setting up news feeds and Google Sheets through IFTTT 172
Setting up your daily personal newsletter 182
Summary 188
Chapter 6: Predict whether Your Content Will Go Viral 189
[ ii ]
Table of Contents
What does research tell us about virality? 190
Sourcing shared counts and content 192
Exploring the features of shareability 195
Exploring image data 196
Clustering 199
Exploring the headlines 202
Exploring the story content 207
Building a predictive content scoring model 210
Evaluating the model 213
Adding new features to our model 216
Summary 218
Chapter 7: Use Machine Learning to Forecast the Stock Market 219
Types of market analysis 221
What does research tell us about the stock market? 221
So, what exactly is a momentum strategy? 222
How to develop a trading strategy 223
Analysis of the data 225
Volatility of the returns 227
Daily returns 229
Statistics for the strategies 230
The mystery strategy 232
Building the regression model 237
Performance of the model 239
Dynamic time warping 246
Evaluating our trades 249
Summary 250
Chapter 8: Classifying Images with Convolutional Neural Networks 251
Image-feature extraction 252
Convolutional neural networks 255
Network topology 256
Convolutional layers and filters 258
Max pooling layers 268
Flattening 270
Fully-connected layers and output 271
Building a convolutional neural network to classify images in the
Zalando Research dataset, using Keras 272
Summary 290
Chapter 9: Building a Chatbot 291
The Turing Test 292
The history of chatbots 293
The design of chatbots 298
Building a chatbot 303
[ iii ]
Table of Contents
Sequence-to-sequence modeling for chatbots 309
Summary 320
Chapter 10: Build a Recommendation Engine 321
Collaborative filtering 322
So, what's collaborative filtering? 322
Predicting the rating for the product 325
Content-based filtering 329
Hybrid systems 330
Collaborative filtering 330
Content-based filtering 330
Building a recommendation engine 331
Summary 348
Chapter 11: What's Next? 349
Summary of the projects 349
Summary 353
Other Books You May Enjoy 354
Index 357
[ iv ]