Machine Learning with R E-Book

Brett Lantz

• Herausgeber: Packt Publishing
• Kategorie: Fachliteratur
• Sprache: Englisch

Seitenzahl: 584

Veröffentlichungsjahr: 2013

Beschreibung

Machine learning, at its core, is concerned with transforming data into actionable knowledge. This fact makes machine learning well-suited to the present-day era of "big data" and "data science". Given the growing prominence of R—a cross-platform, zero-cost statistical programming environment—there has never been a better time to start applying machine learning. Whether you are new to data science or a veteran, machine learning with R offers a powerful set of methods for quickly and easily gaining insight from your data.

"Machine Learning with R" is a practical tutorial that uses hands-on examples to step through real-world application of machine learning. Without shying away from the technical details, we will explore Machine Learning with R using clear and practical examples. Well-suited to machine learning beginners or those with experience. Explore R to find the answer to all of your questions.

How can we use machine learning to transform data into action? Using practical examples, we will explore how to prepare data for analysis, choose a machine learning method, and measure the success of the process.

We will learn how to apply machine learning methods to a variety of common tasks including classification, prediction, forecasting, market basket analysis, and clustering. By applying the most effective machine learning methods to real-world problems, you will gain hands-on experience that will transform the way you think about data.

"Machine Learning with R" will provide you with the analytical tools you need to quickly gain insight from complex data.

Leseprobe

Table of Contents

Machine Learning with R

Credits

About the Author

About the Reviewers

www.PacktPub.com

Support files, eBooks, discount offers and more

Why Subscribe?

Free Access for Packt account holders

Preface

What this book covers

What you need for this book

Who this book is for

Conventions

Reader feedback

Customer support

Downloading the example code

Errata

Piracy

Questions

1. Introducing Machine Learning

The origins of machine learning

Uses and abuses of machine learning

Ethical considerations

How do machines learn?

Abstraction and knowledge representation

Generalization

Assessing the success of learning

Steps to apply machine learning to your data

Choosing a machine learning algorithm

Thinking about the input data

Thinking about types of machine learning algorithms

Matching your data to an appropriate algorithm

Using R for machine learning

Installing and loading R packages

Installing an R package

Installing a package using the point-and-click interface

Loading an R package

Summary

2. Managing and Understanding Data

R data structures

Vectors

Factors

Lists

Data frames

Matrixes and arrays

Managing data with R

Saving and loading R data structures

Importing and saving data from CSV files

Importing data from SQL databases

Exploring and understanding data

Exploring the structure of data

Exploring numeric variables

Measuring the central tendency – mean and median

Measuring spread – quartiles and the five-number summary

Visualizing numeric variables – boxplots

Visualizing numeric variables – histograms

Understanding numeric data – uniform and normal distributions

Measuring spread – variance and standard deviation

Exploring categorical variables

Measuring the central tendency – the mode

Exploring relationships between variables

Visualizing relationships – scatterplots

Examining relationships – two-way cross-tabulations

Summary

3. Lazy Learning – Classification Using Nearest Neighbors

Understanding classification using nearest neighbors

The kNN algorithm

Calculating distance

Choosing an appropriate k

Preparing data for use with kNN

Why is the kNN algorithm lazy?

Diagnosing breast cancer with the kNN algorithm

Step 1 – collecting data

Step 2 – exploring and preparing the data

Transformation – normalizing numeric data

Data preparation – creating training and test datasets

Step 3 – training a model on the data

Step 4 – evaluating model performance

Step 5 – improving model performance

Transformation – z-score standardization

Testing alternative values of k

Summary

4. Probabilistic Learning – Classification Using Naive Bayes

Understanding naive Bayes

Basic concepts of Bayesian methods

Probability

Joint probability

Conditional probability with Bayes' theorem

The naive Bayes algorithm

The naive Bayes classification

The Laplace estimator

Using numeric features with naive Bayes

Example – filtering mobile phone spam with the naive Bayes algorithm

Step 1 – collecting data

Step 2 – exploring and preparing the data

Data preparation – processing text data for analysis

Data preparation – creating training and test datasets

Visualizing text data – word clouds

Data preparation – creating indicator features for frequent words

Step 3 – training a model on the data

Step 4 – evaluating model performance

Step 5 – improving model performance

Summary

5. Divide and Conquer – Classification Using Decision Trees and Rules

Understanding decision trees

Divide and conquer

The C5.0 decision tree algorithm

Choosing the best split

Pruning the decision tree

Example – identifying risky bank loans using C5.0 decision trees

Step 1 – collecting data

Step 2 – exploring and preparing the data

Data preparation – creating random training and test datasets

Step 3 – training a model on the data

Step 4 – evaluating model performance

Step 5 – improving model performance

Boosting the accuracy of decision trees

Making some mistakes more costly than others

Understanding classification rules

Separate and conquer

The One Rule algorithm

The RIPPER algorithm

Rules from decision trees

Example – identifying poisonous mushrooms with rule learners

Step 1 – collecting data

Step 2 – exploring and preparing the data

Step 3 – training a model on the data

Step 4 – evaluating model performance

Step 5 – improving model performance

Summary

6. Forecasting Numeric Data – Regression Methods

Understanding regression

Simple linear regression

Ordinary least squares estimation

Correlations

Multiple linear regression

Example – predicting medical expenses using linear regression

Step 1 – collecting data

Step 2 – exploring and preparing the data

Exploring relationships among features – the correlation matrix

Visualizing relationships among features – the scatterplot matrix

Step 3 – training a model on the data

Step 4 – evaluating model performance

Step 5 – improving model performance

Model specification – adding non-linear relationships

Transformation – converting a numeric variable to a binary indicator

Model specification – adding interaction effects

Putting it all together – an improved regression model

Understanding regression trees and model trees

Adding regression to trees

Example – estimating the quality of wines with regression trees and model trees

Step 1 – collecting data

Step 2 – exploring and preparing the data

Step 3 – training a model on the data

Visualizing decision trees

Step 4 – evaluating model performance

Measuring performance with mean absolute error

Step 5 – improving model performance

Summary

7. Black Box Methods – Neural Networks and Support Vector Machines

Understanding neural networks

From biological to artificial neurons

Activation functions

Network topology

The number of layers

The direction of information travel

The number of nodes in each layer

Training neural networks with backpropagation

Modeling the strength of concrete with ANNs

Step 1 – collecting data

Step 2 – exploring and preparing the data

Step 3 – training a model on the data

Step 4 – evaluating model performance

Step 5 – improving model performance

Understanding Support Vector Machines

Classification with hyperplanes

Finding the maximum margin

The case of linearly separable data

The case of non-linearly separable data

Using kernels for non-linear spaces

Performing OCR with SVMs

Step 1 – collecting data

Step 2 – exploring and preparing the data

Step 3 – training a model on the data

Step 4 – evaluating model performance

Step 5 – improving model performance

Summary

8. Finding Patterns – Market Basket Analysis Using Association Rules

Understanding association rules

The Apriori algorithm for association rule learning

Measuring rule interest – support and confidence

Building a set of rules with the Apriori principle

Example – identifying frequently purchased groceries with association rules

Step 1 – collecting data

Step 2 – exploring and preparing the data

Data preparation – creating a sparse matrix for transaction data

Visualizing item support – item frequency plots

Visualizing transaction data – plotting the sparse matrix

Step 3 – training a model on the data

Step 4 – evaluating model performance

Step 5 – improving model performance

Sorting the set of association rules

Taking subsets of association rules

Saving association rules to a file or data frame

Summary

9. Finding Groups of Data – Clustering with k-means

Understanding clustering

Clustering as a machine learning task

The k-means algorithm for clustering

Using distance to assign and update clusters

Choosing the appropriate number of clusters

Finding teen market segments using k-means clustering

Step 1 – collecting data

Step 2 – exploring and preparing the data

Data preparation – dummy coding missing values

Data preparation – imputing missing values

Step 3 – training a model on the data

Step 4 – evaluating model performance

Step 5 – improving model performance

Summary

10. Evaluating Model Performance

Measuring performance for classification

Working with classification prediction data in R

A closer look at confusion matrices

Using confusion matrices to measure performance

Beyond accuracy – other measures of performance

The kappa statistic

Sensitivity and specificity

Precision and recall

The F-measure

Visualizing performance tradeoffs

ROC curves

Estimating future performance

The holdout method

Cross-validation

Bootstrap sampling

Summary

11. Improving Model Performance

Tuning stock models for better performance

Using caret for automated parameter tuning

Creating a simple tuned model

Customizing the tuning process

Improving model performance with meta-learning

Understanding ensembles

Bagging

Boosting

Random forests

Training random forests

Evaluating random forest performance

Summary

12. Specialized Machine Learning Topics

Working with specialized data

Getting data from the Web with the RCurl package

Reading and writing XML with the XML package

Reading and writing JSON with the rjson package

Reading and writing Microsoft Excel spreadsheets using xlsx

Working with bioinformatics data

Working with social network data and graph data

Improving the performance of R

Managing very large datasets

Making data frames faster with data.table

Creating disk-based data frames with ff

Using massive matrices with bigmemory

Learning faster with parallel computing

Measuring execution time

Working in parallel with foreach

Using a multitasking operating system with multicore

Networking multiple workstations with snow and snowfall

Parallel cloud computing with MapReduce and Hadoop

GPU computing

Deploying optimized learning algorithms

Building bigger regression models with biglm

Growing bigger and faster random forests with bigrf

Training and evaluating models in parallel with caret

Summary

Index

Machine Learning with R

Machine Learning with R

Copyright © 2013 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 author, nor Packt Publishing, and its dealers and distributors will be held liable for any damages caused or alleged to be 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.

First published: October 2013

Production Reference: 1211013

Published by Packt Publishing Ltd.

Livery Place

35 Livery Street

Birmingham B3 2PB, UK.

ISBN 978-1-78216-214-8

www.packtpub.com

Cover Image by Abhishek Pandey (<[email protected]>)

Credits

Author

Brett Lantz

Reviewers

Jia Liu

Mzabalazo Z. Ngwenya

Abhinav Upadhyay

Acquisition Editor

James Jones

Lead Technical Editor

Azharuddin Sheikh

Technical Editors

Pooja Arondekar

Pratik More

Anusri Ramchandran

Harshad Vairat

Project Coordinator

Anugya Khurana

Proofreaders

Simran Bhogal

Ameesha Green

Paul Hindle

Indexer

Tejal Soni

Graphics

Ronak Dhruv

Production Coordinator

Nilesh R. Mohite

Cover Work

Nilesh R. Mohite

About the Author

Brett Lantz has spent the past 10 years using innovative data methods to understand human behavior. A sociologist by training, he was first enchanted by machine learning while studying a large database of teenagers' social networking website profiles. Since then, he has worked on interdisciplinary studies of cellular telephone calls, medical billing data, and philanthropic activity, among others. When he's not spending time with family, following college sports, or being entertained by his dachshunds, he maintains dataspelunking.com, a website dedicated to sharing knowledge about the search for insight in data.

This book could not have been written without the support of my family and friends. In particular, my wife Jessica deserves many thanks for her patience and encouragement throughout the past year. My son Will (who was born while Chapter 10 was underway), also deserves special mention for his role in the writing process; without his gracious ability to sleep through the night, I could not have strung together a coherent sentence the next morning. I dedicate this book to him in the hope that one day he is inspired to follow his curiosity wherever it may lead.

I am also indebted to many others who supported this book indirectly. My interactions with educators, peers, and collaborators at the University of Michigan, the University of Notre Dame, and the University of Central Florida seeded many of the ideas I attempted to express in the text. Additionally, without the work of researchers who shared their expertise in publications, lectures, and source code, this book might not exist at all. Finally, I appreciate the efforts of the R team and all those who have contributed to R packages, whose work ultimately brought machine learning to the masses.

About the Reviewers

Jia Liu holds a Master's degree in Statistics from the University of Maryland, Baltimore County, and is presently a PhD candidate in statistics from Iowa State University. Her research interests include mixed-effects model, Bayesian method, Boostrap method, reliability, design of experiments, machine learning and data mining. She has two year's experience as a student consultant in statistics and two year's internship experience in agriculture and pharmaceutical industry.

Mzabalazo Z. Ngwenya has worked extensively in the field of statistical consulting and currently works as a biometrician. He holds an MSc in Mathematical Statistics from the University of Cape Town and is at present studying for a PhD (at the School of Information Technology, University of Pretoria), in the field of Computational Intelligence. His research interests include statistical computing, machine learning, and spatial statistics. Previously, he was involved in reviewing Learning RStudio for R Statistical Computing (Van de Loo and de Jong, 2012), and R Statistical Application Development by Example beginner's guide (Prabhanjan Narayanachar Tattar , 2013).

Abhinav Upadhyay finished his Bachelor's degree in 2011 with a major in Information Technology. His main areas of interest include machine learning and information retrieval.

In 2011, he worked for the NetBSD Foundation as part of the Google Summer of Code program. During that period, he wrote a search engine for Unix manual pages. This project resulted in a new implementation of the apropos utility for NetBSD.

Currently, he is working as a Development Engineer for SocialTwist. His day-to-day work involves writing system level tools and frameworks to manage the product infrastructure.

He is also an open source enthusiast and quite active in the community. In his free time, he maintains and contributes to several open source projects.

www.PacktPub.com

Support files, eBooks, discount offers and more

You might want to visit www.PacktPub.com for support files and downloads related to your book.

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.PacktPub.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.PacktPub.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.

http://PacktLib.PacktPub.com

Do you need instant solutions to your IT questions? PacktLib is Packt's online digital book library. Here, you can access, read and search across Packt's entire library of books.

Why Subscribe?

Fully searchable across every book published by PacktCopy and paste, print and bookmark contentOn demand and accessible via web browser

Free Access for Packt account holders

If you have an account with Packt at www.PacktPub.com, you can use this to access PacktLib today and view nine entirely free books. Simply use your login credentials for immediate access.

Preface

Machine learning, at its core, is concerned with algorithms that transform information into actionable intelligence. This fact makes machine learning well-suited to the present day era of Big Data. Without machine learning, it would be nearly impossible to keep up with the massive stream of information.

Given the growing prominence of R—a cross-platform, zero-cost statistical programming environment—there has never been a better time to start using machine learning. R offers a powerful but easy-to-learn set of tools that can assist you with finding data insights.

By combining hands-on case studies with the essential theory that you need to understand how things work under the hood, this book provides all the knowledge that you will need to start applying machine learning to your own projects.

What this book covers

Chapter 1, Introducing Machine Learning, presents the terminology and concepts that define and distinguish machine learners, as well as a method for matching a learning task with the appropriate algorithm.

Chapter 2, Managing and Understanding Data, provides an opportunity to get your hands dirty working with data in R. Essential data structures and procedures used for loading, exploring, and understanding data are discussed.

Chapter 3, Lazy Learning – Classification Using Nearest Neighbors, teaches you how to understand and apply a simple yet powerful learning algorithm to your first machine learning task: identifying malignant samples of cancer.

Chapter 4, Probabilistic Learning – Classification Using Naive Bayes, reveals the essential concepts of probability that are used in cutting-edge spam filtering systems. You'll learn the basics of text mining in the process of building your own spam filter.

Chapter 5, Divide and Conquer – Classification Using Decision Trees and Rules, explores a couple of learning algorithms whose predictions are not only accurate but easily explained. We'll apply these methods to tasks where transparency is important.

Chapter 6, Forecasting Numeric Data – Regression Methods, introduces machine learning algorithms used for making numeric predictions. As these techniques are heavily embedded in the field of statistics, you will also learn the essential metrics needed to make sense of numeric relationships.

Chapter 7, Black Box Methods – Neural Networks and Support Vector Machines, covers two extremely complex yet powerful machine learning algorithms. Though the mathematics may appear intimidating, we will work through examples that illustrate their inner workings in simple terms.

Chapter 8, Finding Patterns – Market Basket Analysis Using Association Rules, exposes the algorithm for the recommendation systems used at many retailers. If you've ever wondered how retailers seem to know your purchasing habits better than you know them yourself, this chapter will reveal their secrets.

Chapter 9, Finding Groups of Data – Clustering with k-means, is devoted to a procedure that locates clusters of related items. We'll utilize this algorithm to identify segments of profiles within a web-based community.

Chapter 10, Evaluating Model Performance, provides information on measuring the success of a machine learning project, and obtaining a reliable estimate of the learner's performance on future data.

Chapter 11, Improving Model Performance, reveals the methods employed by the teams found at the top of machine learning competition leader boards. If you have a competitive streak, or simply want to get the most out of your data, you'll need to add these techniques to your repertoire.

Chapter 12, Specialized Machine Learning Topics, explores the frontiers of machine learning. From working with Big Data to making R work faster, the topics covered will help you push the boundaries of what is possible with R.

What you need for this book

The examples in this book were written for and tested with R Version 2.15.3 on both Microsoft Windows and Mac OS X, though they are likely to work with any recent version of R.

Who this book is for

This book is intended for anybody hoping to use data for action. Perhaps you already know a bit about machine learning, but have never used R; or perhaps you know a little R but are new to machine learning. In any case, this book will get you up and running quickly. It would be helpful to have a bit of familiarity with basic math and programming concepts, but no prior experience is required. You need only curiosity.

Conventions

In this book, you will find a number of styles of text that distinguish between different kinds of information. Here are some examples of these styles, and an explanation of their meaning.

Code words in text, database table names, folder names, filenames, file extensions, pathnames, dummy URLs, user input, and Twitter handles are shown as follows: "To fit a linear regression model to data with R, the lm() function can be used."

Any command-line input or output is written as follows:

> pairs.panels(insurance[c("age", "bmi", "children", "charges")])

New terms and important words are shown in bold. Words that you see on the screen, in menus or dialog boxes for example, appear in the text like this: "Instead, ham messages use words such as can, sorry, need, and time."

Note

Warnings or important notes appear in a box like this.

Tip

Tips and tricks appear like this.

Reader feedback

Feedback from our readers is always welcome. Let us know what you think about this book—what you liked or may have disliked. Reader feedback is important for us to develop titles that you really get the most out of.

To send us general feedback, simply send an e-mail to <[email protected]>, and mention the book title via the subject of your message.

If there is a topic that you have expertise in and you are interested in either writing or contributing to a book, see our author guide on www.packtpub.com/authors.

Customer support

Now that you are the proud owner of a Packt book, we have a number of things to help you to get the most from your purchase.

Downloading the example code

You can download the example code files for all Packt books you have purchased from your account at http://www.packtpub.com. If you purchased this book elsewhere, you can visit http://www.packtpub.com/support and register to have the files e-mailed directly to you.

Errata

Although we have taken every care to ensure the accuracy of our content, mistakes do happen. If you find a mistake in one of our books—maybe a mistake in the text or the code—we would be grateful if you would report this to us. By doing so, you can save other readers from frustration and help us improve subsequent versions of this book. If you find any errata, please report them by visiting http://www.packtpub.com/submit-errata, selecting your book, clicking on the erratasubmissionform link, and entering the details of your errata. Once your errata are verified, your submission will be accepted and the errata will be uploaded on our website, or added to any list of existing errata, under the Errata section of that title. Any existing errata can be viewed by selecting your title from http://www.packtpub.com/support.

Piracy

Piracy of copyright material on the Internet is an ongoing problem across all media. At Packt, we take the protection of our copyright and licenses very seriously. If you come across any illegal copies of our works, in any form, on the Internet, please provide us with the location address or website name immediately so that we can pursue a remedy.

Please contact us at <[email protected]> with a link to the suspected pirated material.

We appreciate your help in protecting our authors, and our ability to bring you valuable content.

Questions

You can contact us at <[email protected]> if you are having a problem with any aspect of the book, and we will do our best to address it.

Chapter 1. Introducing Machine Learning

If science fiction stories are to be believed, teaching machines to learn will inevitably lead to apocalyptic wars between machines and their makers. In the early stages, computers are taught to play simple games of tic-tac-toe and chess. Later, machines are given control of traffic lights and communications, followed by military drones and missiles. The machines' evolution takes an ominous turn once the computers become sentient and learn how to teach themselves. Having no more need for human programmers, humankind is then "deleted."

Thankfully, at the time of this writing, machines still require user input.

Your impressions of machine learning may be very heavily influenced by these types of mass media depictions of artificial intelligence. And even though there may be a hint of truth to such tales; in reality, machine learning is focused on more practical applications. The task of teaching a computer to learn is tied more closely to a specific problem that would be a computer that can play games, ponder philosophy, or answer trivial questions. Machine learning is more like training an employee than raising a child.

Putting these stereotypes aside, by the end of this chapter, you will have gained a far more nuanced understanding of machine learning. You will be introduced to the fundamental concepts that define and differentiate the most commonly used machine learning approaches.

You will learn:

The origins and practical applications of machine learningHow knowledge is defined and represented by computersThe basic concepts that differentiate machine learning approaches

In a single sentence, you could say that machine learning provides a set of tools that use computers to transform data into actionable knowledge. To learn more about how the process works, read on.

The origins of machine learning

Since birth, we are inundated with data. Our body's sensors—the eyes, ears, nose, tongue, and nerves—are continually assailed with raw data that our brain translates into sights, sounds, smells, tastes, and textures. Using language, we are able to share these experiences with others.

The earliest databases recorded information from the observable environment. Astronomers recorded patterns of planets and stars; biologists noted results from experiments crossbreeding plants and animals; and cities recorded tax payments, disease outbreaks, and populations. Each of these required a human being to first observe and second, record the observation. Today, such observations are increasingly automated and recorded systematically in ever-growing computerized databases.

The invention of electronic sensors has additionally contributed to an increase in the richness of recorded data. Specialized sensors see, hear, smell, or taste. These sensors process the data far differently than a human being would, and in many ways, this is a benefit. Without the need for translation into human language, the raw sensory data remains objective.

Tip

It is important to note that although a sensor does not have a subjective component to its observations, it does not necessarily report truth (if such a concept can be defined). A camera taking photographs in black and white might provide a far different depiction of its environment than one shooting pictures in color. Similarly, a microscope provides a far different depiction of reality than a telescope.

Between databases and sensors, many aspects of our lives are recorded. Governments, businesses, and individuals are recording and reporting all manners of information from the monumental to the mundane. Weather sensors record temperature and pressure data, surveillance cameras watch sidewalks and subway tunnels, and all manner of electronic behaviors are monitored: transactions, communications, friendships, and many others.

This deluge of data has led some to state that we have entered an era of Big Data, but this may be a bit of a misnomer. Human beings have always been surrounded by data. What makes the current era unique is that we have easy data. Larger and more interesting data sets are increasingly accessible through the tips of our fingers, only a web search away. We now live in a period with vast quantities of data that can be directly processed by machines. Much of this information has the potential to inform decision making, if only there was a systematic way of making sense from it all.

The field of study interested in the development of computer algorithms for transforming data into intelligent action is known as machine learning. This field originated in an environment where the available data, statistical methods, and computing power rapidly and simultaneously evolved. Growth in data necessitated additional computing power, which in turn spurred the development of statistical methods for analyzing large datasets. This created a cycle of advancement allowing even larger and more interesting data to be collected.

A closely related sibling of machine learning, data mining, is concerned with the generation of novel insight from large databases (not to be confused with the pejorative term "data mining," describing the practice of cherry-picking data to support a theory). Although there is some disagreement over how widely the two fields overlap, a potential point of distinction is that machine learning tends to be focused on performing a known task, whereas data mining is about the search for hidden nuggets of information. For instance, you might use machine learning to teach a robot to drive a car, whereas you would utilize data mining to learn what type of cars are the safest.

Tip

Machine learning algorithms are virtually a prerequisite for data mining but the opposite is not true. In other words, you can apply machine learning to tasks that do not involve data mining, but if you are using data mining methods, you are almost certainly using machine learning.

Uses and abuses of machine learning

At its core, machine learning is primarily interested in making sense of complex data. This is a broadly applicable mission, and largely application agnostic. As you might expect, machine learning is used widely. For instance, it has been used to:

Predict the outcomes of electionsIdentify and filter spam messages from e-mailForesee criminal activityAutomate traffic signals according to road conditionsProduce financial estimates of storms and natural disastersExamine customer churnCreate auto-piloting planes and auto-driving carsIdentify individuals with the capacity to donateTarget advertising to specific types of consumers

For now, don't worry about exactly how the machines learn to perform these tasks; we will get into the specifics later. But across each of these contexts, the process is the same. A machine learning algorithm takes data and identifies patterns that can be used for action. In some cases, the results are so successful that they seem to reach near-legendary status.

One possibly apocryphal tale is of a large retailer in the United States, which employed machine learning to identify expectant mothers for targeted coupon mailings. If mothers-to-be were targeted with substantial discounts, the retailer hoped they would become loyal customers who would then continue to purchase profitable items like diapers, formula, and toys.

By applying machine learning methods to purchase data, the retailer believed it had learned some useful patterns. Certain items, such as prenatal vitamins, lotions, and washcloths could be used to identify with a high degree of certainty not only whether a woman was pregnant, but also when the baby was due.

After using this data for a promotional mailing, an angry man contacted the retailer and demanded to know why his teenage daughter was receiving coupons for maternity items. He was furious that the merchant seemed to be encouraging teenage pregnancy. Later on, as a manager called to offer an apology, it was the father that ultimately apologized; after confronting his daughter, he had discovered that she was indeed pregnant.

Whether completely true or not, there is certainly an element of truth to the preceding tale. Retailers, do in fact, routinely analyze their customers' transaction data. If you've ever used a shopper's loyalty card at your grocer, coffee shop, or another retailer, it is likely that your purchase data is being used for machine learning.

Retailers use machine learning methods for advertising, targeted promotions, inventory management, or the layout of the items in the store. Some retailers have even equipped checkout lanes with devices that print coupons for promotions based on the items in the current transaction. Websites also routinely do this to serve advertisements based on your web browsing history. Given the data from many individuals, a machine learning algorithm learns typical patterns of behavior that can then be used to make recommendations.

Despite being familiar with the machine learning methods working behind the scenes, it still feels a bit like magic when a retailer or website seems to know me better than I know myself. Others may be less thrilled to discover that their data is being used in this manner. Therefore, any person wishing to utilize machine learning or data mining would be remiss not to at least briefly consider the ethical implications of the art.

Ethical considerations

Due to the relative youth of machine learning as a discipline and the speed at which it is progressing, the associated legal issues and social norms are often quite uncertain and constantly in flux. Caution should be exercised when obtaining or analyzing data in order to avoid breaking laws, violating terms of service or data use agreements, abusing the trust, or violating privacy of the customers or the public.

Tip

The informal corporate motto of Google, an organization, which collects perhaps more data on individuals than any other, is "don't be evil." This may serve as a reasonable starting point for forming your own ethical guidelines, but it may not be sufficient.

Certain jurisdictions may prevent you from using racial, ethnic, religious, or other protected class data for business reasons, but keep in mind that excluding this data from your analysis may not be enough—machine learning algorithms might inadvertently learn this information independently. For instance, if a certain segment of people generally live in a certain region, buy a certain product, or otherwise behave in a way that uniquely identifies them as a group, some machine learning algorithms can infer the protected information from seemingly innocuous data. In such cases, you may need to fully "de-identify" these people by excluding any potentially identifying data in addition to the protected information.

Apart from the legal consequences, using data inappropriately may hurt your bottom line. Customers may feel uncomfortable or become spooked if aspects of their lives they consider private are made public. Recently, several high-profile web applications have experienced a mass exodus of users who felt exploited when the applications' terms of service agreements changed and their data was used for purposes beyond what the users had originally agreed upon. The fact that privacy expectations differ by context, by age cohort, and by locale, adds complexity to deciding the appropriate use of personal data. It would be wise to consider the cultural implications of your work before you begin on your project.

Tip

The fact that you can use data for a particular end does not always mean that you should.

How do machines learn?

A commonly cited formal definition of machine learning, proposed by computer scientist Tom M. Mitchell, says that a machine is said to learn if it is able to take experience and utilize it such that its performance improves up on similar experiences in the future. This definition is fairly exact, yet says little about how machine learning techniques actually learn to transform data into actionable knowledge.

Tip

Although it is not strictly necessary to understand the theoretical basis of machine learning prior to using it, this foundation provides an insight into the distinctions among machine learning algorithms. Because machine learning algorithms are modeled in many ways on human minds, you may even discover yourself examining your own mind in a different light.

Regardless of whether the learner is a human or a machine, the basic learning process is similar. It can be divided into three components as follows:

Data input: It utilizes observation, memory storage, and recall to provide a factual basis for further reasoning.Abstraction: It involves the translation of data into broader representations.Generalization: It uses abstracted data to form a basis for action.

To better understand the learning process, think about the last time you studied for a difficult test, perhaps for a university final exam or a career certification. Did you wish for an eidetic (that is, photographic) memory? If so, you may be disappointed to learn that perfect recall is unlikely to save you much effort. Without a higher understanding, your knowledge is limited exactly to the data input, meaning only what you had seen before and nothing more. Therefore, without knowledge of all the questions that could appear on the exam, you would be stuck attempting to memorize answers to every question that could conceivably be asked. Obviously, this is an unsustainable strategy.

Instead, a better strategy is to spend time selectively managing only a smaller set of key ideas. The commonly used learning strategies of creating an outline or a concept map are similar to how a machine performs knowledge abstraction. The tools define relationships among information and in doing so, depict difficult ideas without needing to memorize them word-for-word. It is a more advanced form of learning because it requires that the learner puts the topic into his or her own words.

It is always a tense moment when the exam is graded and the learning strategies are either vindicated or implicated with a high or low mark. Here, one discovers whether the learning strategies generalized to the questions that the teacher or professor had selected. Generalization requires a breadth of abstracted data, as well as a higher-level understanding of how to apply such knowledge to unforeseen topics. A good teacher can be quite helpful in this regard.

Keep in mind that although we have illustrated the learning process as three distinct steps, they are merely organized this way for illustrative purposes. In reality, the three components of learning are inextricably linked. In particular, the stages of abstraction and generalization are so closely related that it would be impossible to perform one without the other. In human beings, the entire process happens subconsciously. We recollect, deduce, induct, and intuit. Yet for a computer, these processes must be made explicit. On the other hand, this is a benefit of machine learning. Because the process is transparent, the learned knowledge can be examined and utilized for future action.

Abstraction and knowledge representation

Representing raw input data in a structured format is the quintessential task for a learning algorithm. Prior to this point, the data is merely ones and zeros on a disk or in memory; they have no meaning. The work of assigning a meaning to data occurs during the abstraction process.

The connection between ideas and reality is exemplified by the famous René Magritte painting The Treachery of Images shown as follows:

Source: http://collections.lacma.org/node/239578

The painting depicts a tobacco pipe with the caption Ceci n'est pas une pipe ("this is not a pipe"). The point Magritte was illustrating is that a representation of a pipe is not truly a pipe. In spite of the fact that the pipe is not real, anybody viewing the painting easily recognizes that the picture is a pipe, suggesting that observers' minds are able to connect the picture of a pipe to the idea of a pipe, which can then be connected to an actual pipe that could be held in the hand. Abstracted connections like this are the basis of knowledge representation, the formation of logical structures that assist with turning raw sensory information into a meaningful insight.

During the process of knowledge representation, the computer summarizes raw inputs in amodel, an explicit description of the structured patterns among data. There are many different types of models. You may already be familiar with some. Examples include:

EquationsDiagrams such as trees and graphsLogical if/else rulesGroupings of data known as clusters

The choice of model is typically not left up to the machine. Instead, the model is dictated by the learning task and the type of data being analyzed. Later in this chapter, we will discuss methods for choosing the type of model in more detail.

The process of fitting a particular model to a dataset is known as training. Why is this not called learning? First, note that the learning process does not end with the step of data abstraction. Learning requires an additional step to generalize the knowledge to future data. Second, the term training more accurately describes the actual process undertaken when the model is fitted to the data. Learning implies a sort of inductive, bottom-up reasoning. Training better connotes the fact that the machine learning model is imposed by the human teacher onto the machine student, providing the computer with a structure it attempts to model after.

When the model has been trained, the data has been transformed into an abstract form that summarizes the original information. It is important to note that the model does not itself provide additional data, yet it is sometimes interesting on its own. How can this be? The reason is that by imposing an assumed structure on the underlying data, it gives insight into the unseen and provides a theory about how the data is related. Take for instance the discovery of gravity. By fitting equations to observational data, Sir Isaac Newton deduced the concept of gravity. But gravity was always present. It simply wasn't recognized as a concept until the model noted it in abstract terms—specifically, by becoming the g term in a model that explains observations of falling objects.

Most models will not result in the development of theories that shake up scientific thought for centuries. Still, your model might result in the discovery of previously unseen relationships among data. A model trained on genomic data might find several genes that when combined are responsible for the onset of diabetes; banks might discover a seemingly innocuous type of transaction that systematically appears prior to fraudulent activity; psychologists might identify a combination of characteristics indicating a new disorder. The underlying relationships were always present; but in conceptualizing the information in a different format, a model presents the connections in a new light.

Generalization

Recall that the learning process is not complete until the learner is able to use its abstracted knowledge for future action. Yet an issue remains before the learner can proceed—there are countless underlying relationships that might be identified during the abstraction process and myriad ways to model these relationships. Unless the number of potential theories is limited, the learner will be unable to utilize the information. It would be stuck where it started, with a large pool of information but no actionable insight.

The term generalization describes the process of turning abstracted knowledge into a form that can be utilized for action. Generalization is a somewhat vague process that is a bit difficult to describe. Traditionally, it has been imagined as a search through the entire set of models (that is, theories) that could have been abstracted during training. Specifically, if you imagine a hypothetical set containing every possible theory that could be established from the data, generalization involves the reduction of this set into a manageable number of important findings.

Generally, it is not feasible to reduce the number of potential concepts by examining them one-by-one and determining which are the most useful. Instead, machine learning algorithms generally employ shortcuts that more quickly divide the set of concepts. Toward this end, the algorithm will employ heuristics, or educated guesses about the where to find the most important concepts.

Tip

Because the heuristics utilize approximations and other rules of thumb, they are not guaranteed to find the optimal set of concepts that model the data. However, without utilizing these shortcuts, finding useful information in a large dataset would be infeasible.

Heuristics are routinely used by human beings to quickly generalize experience to new scenarios. If you have ever utilized gut instinct to make a snap decision prior to fully evaluating your circumstances, you were intuitively using mental heuristics.

For example, the availability heuristic is the tendency for people to estimate the likelihood of an event by how easily examples can be recalled. The availability heuristic might help explain the prevalence of the fear of airline travel relative to automobile travel, despite automobiles being statistically more dangerous. Accidents involving air travel are highly publicized and traumatic events, and are likely to be very easily recalled, whereas car accidents barely warrant a mention in the newspaper.

The preceding example illustrates the potential for heuristics to result in illogical conclusions. Browsing a list of common logical fallacies, one is likely to note many that seem rooted in heuristic-based thinking. For instance, the gambler's fallacy, or the belief that a run of bad luck implies that a stretch of better luck is due, may be resultant from the application of the representativeness heuristic, which erroneously led the gambler to believe that all random sequences are balanced since most random sequences are balanced.

The folly of misapplied heuristics is not limited to human beings. The heuristics employed by machine learning algorithms also sometimes result in erroneous conclusions. If the conclusions are systematically imprecise, the algorithm is said to have a bias. For example, suppose that a machine learning algorithm learned to identify faces by finding two circles, or eyes, positioned side-by-side above a line for a mouth. The algorithm might then have trouble with, or be biased against faces that do not conform to its model. This may include faces with glasses, turned at an angle, looking sideways, or with darker skin tones. Similarly, it could be biased toward faces with lighter eye colors or other characteristics that do not conform to its understanding of the world.

In modern usage, the word bias has come to carry quite negative connotations. Various forms of media frequently claim to be free from bias, and claim to report the facts objectively, untainted by emotion. Still, consider for a moment the possibility that a little bias might be useful. Without a bit of arbitrariness, might it be a bit difficult to decide among several competing choices, each with distinct strengths and weaknesses? Indeed, some recent studies in the field of psychology have suggested that individuals born with damage to portions of the brain responsible for emotion are ineffectual at decision making, and might spend hours debating simple decisions such as what color shirt to wear or where to eat lunch. Paradoxically, bias is what blinds us from some information while also allowing us to utilize other information for action.

Assessing the success of learning

Bias is a necessary evil associated with the abstraction and generalization process inherent in any machine learning task. Every learner has its weaknesses and is biased in a particular way; there is no single model to rule them all. Therefore, the final step in the generalization process is to determine the model's success in spite of its biases.

After a model has been trained on an initial dataset, the model is tested on a new dataset, and judged on how well its characterization of the training data generalizes to the new data. It's worth noting that it is exceedingly rare for a model to perfectly generalize to every unforeseen case.

In part, the failure for models to perfectly generalize is due to the problem of noise, or unexplained variations in data. Noisy data is caused by seemingly random events, such as:

Measurement error due to imprecise sensors that sometimes add or subtract a bit from the readingIssues with reporting data, such as respondents reporting random answers to survey questions in order to finish more quicklyErrors caused when data is recorded incorrectly, including missing, null, truncated, incorrectly coded, or corrupted values

Trying to model the noise in data is the basis of a problem calledoverfitting. Because noise is unexplainable by definition, attempting to explain the noise will result in erroneous conclusions that do not generalize well to new cases. Attempting to generate theories to explain the noise also results in more complex models that are more likely to ignore the true pattern the learner is trying to identify. A model that seems to perform well during training but does poorly during testing is said to be overfitted to the training dataset as it does not generalize well.

Solutions to the problem of overfitting are specific to particular machine learning approaches. For now, the important point is to be aware of the issue. How well models are able to handle noisy data is an important source of distinction among them.

Steps to apply machine learning to your data

Any machine learning task can be broken down into a series of more manageable steps. This book has been organized according to the following process:

Collecting data: Whether the data is written on paper, recorded in text files and spreadsheets, or stored in an SQL database, you will need to gather it in an electronic format suitable for analysis. This data will serve as the learning material an algorithm uses to generate actionable knowledge.Exploring and preparing the data: The quality of any machine learning project is based largely on the quality of data it uses. This step in the machine learning process tends to require a great deal of human intervention. An often cited statistic suggests that 80 percent of the effort in machine learning is devoted to data. Much of this time is spent learning more about the data and its nuances during a practice called data exploration.Training a model on the data: By the time the data has been prepared for analysis, you are likely to have a sense of what you are hoping to learn from the data. The specific machine learning task will inform the selection of an appropriate algorithm, and the algorithm will represent the data in the form of a model.Evaluating model performance: Because each machine learning model results in a biased solution to the learning problem, it is important to evaluate how well the algorithm learned from its experience. Depending on the type of model used, you might be able to evaluate the accuracy of the model using a test dataset, or you may need to develop measures of performance specific to the intended application.Improving model performance: If better performance is needed, it becomes necessary to utilize more advanced strategies to augment the performance of the model. Sometimes, it may be necessary to switch to a different type of model altogether. You may need to supplement your data with additional data, or perform additional preparatory work as in step two of this process.

After these steps have been completed, if the model appears to be performing satisfactorily, it can be deployed for its intended task. As the case may be, you might utilize your model to provide score data for predictions (possibly in real time), for projections of financial data, to generate useful insight for marketing or research, or to automate tasks such as mail delivery or flying aircraft. The successes and failures of the deployed model might even provide additional data to train the next generation of your model.

Choosing a machine learning algorithm

The process of choosing a machine learning algorithm involves matching the characteristics of the data to be learned to the biases of the available approaches. Since the choice of a machine learning algorithm is largely dependent upon the type of data you are analyzing and the proposed task at hand, it is often helpful to be thinking about this process while you are gathering, exploring, and cleaning your data.

Tip

It may be tempting to learn a couple of machine learning techniques and apply them to everything, but resist this temptation. No machine learning approach is best for every circumstance. This fact is described by the No Free Lunch theorem, introduced by David Wolpert in 1996. For more information, visit: http://www.no-free-lunch.org.

Thinking about the input data

All machine learning algorithms require input training data. The exact format may differ, but in its most basic form, input data takes the form of examples and features.

An example is literally a single exemplary instance of the underlying concept to be learned; it is one set of data describing the atomic unit of interest for the analysis. If you were building a learning algorithm to identify spam e-mail, the examples would be data from many individual electronic messages. To detect cancerous tumors, the examples might comprise biopsies from a number of patients.

The phrase unit of observation is used to describe the units that the examples are measured in. Commonly, the unit of observation is in the form of transactions, persons, time points, geographic regions, or measurements. Other possibilities include combinations of these such as person years, which would denote cases where the same person is tracked over multiple time points.

A feature is a characteristic or attribute of an example, which might be useful for learning the desired concept. In the previous examples, attributes in the spam detection dataset might consist of the words used in the e-mail messages. For the cancer dataset, the attributes might be genomic data from the biopsied cells, or measured characteristics of the patient such as weight, height, or blood pressure.

The following spreadsheet shows a dataset in matrix format, which means that each example has the same number of features. In matrix data, each row in the spreadsheet is an example and each column is a feature. Here, the rows indicate examples of automobiles while the columns record various features of the cars such as the price, mileage, color, and transmission. Matrix format data is by far the most common form used in machine learning, though as you will see in later chapters, other forms are used occasionally in specialized cases.

Features come in various forms as well. If a feature represents a characteristic measured in numbers, it is unsurprisingly called numeric. Alternatively, if it measures an attribute that is represented by a set of categories, the feature is called categorical or nominal. A special case of categorical variables is called ordinal, which designates a nominal variable with categories falling in an ordered list. Some examples of ordinal variables include clothing sizes such as small, medium, and large, or a measurement of customer satisfaction on a scale from 1 to 5. It is important to consider what the features represent because the type and number of features in your dataset will assist with determining an appropriate machine learning algorithm for your task.

Thinking about types of machine learning algorithms

Machine learning algorithms can be divided into two main groups: supervised learners that are used to construct predictive models, and unsupervised learners that are used to build descriptive models. Which type you will need to use depends on the learning task you hope to accomplish.

A predictive model is used for tasks that involve, as the name implies, the prediction of one value using other values in the dataset. The learning algorithm attempts to discover and model the relationship among the target feature (the feature being predicted) and the other features. Despite the common use of the word "prediction" to imply forecasting predictive models need not necessarily foresee future events. For instance, a predictive model could be used to predict past events such as the date of a baby's conception using the mother's hormone levels; or, predictive models could be used in real time to control traffic lights during rush hours.

Because predictive models are given clear instruction on what they need to learn and how they are intended to learn it, the process of training a predictive model is known as supervised learning. The supervision does not refer to human involvement, but rather the fact that the target values provide a supervisory role, which indicates to the learner the task it needs to learn. Specifically, given a set of data, the learning algorithm attempts to optimize a function (the model) to find the combination of feature values that result in the target output.

The often used supervised machine learning task of predicting which category an example belongs to is known as classification. It is easy to think of potential uses for a classifier. For instance, you could predict whether:

A football team will win or loseA person will live past the age of 100An applicant will default on a loanAn earthquake will strike next year

The target feature to be predicted is a categorical feature known as the class and is divided into categories called levels. A class can have two or more levels, and the levels need not necessarily be ordinal. Because classification is so widely used in machine learning, there are many types of classification algorithms.

Supervised learners can also be used to predict numeric data such as income, laboratory values, test scores, or counts of items. To predict such numeric values, a common form of numeric prediction fits linear regression models to the input data. Although regression models are not the only type of numeric models, they are by far the most widely used. Regression methods are widely used for forecasting, as they quantify in exact terms the association between the inputs and the target, including both the magnitude and uncertainty of the relationship.

Tip

Since it is easy to convert numbers to categories (for example, ages 13 to 19 are teenagers) and categories to numbers (for example, assign 1 to all males, 0 to all females), the boundary between classification models and numeric prediction models is not necessarily firm.

A descriptive model is used for tasks that would benefit from the insight gained from summarizing data in new and interesting ways. As opposed to predictive models that predict a target of interest; in a descriptive model, no single feature is more important than any other. In fact, because there is no target to learn, the process of training a descriptive model is called unsupervised learning. Although it can be more difficult to think of applications for descriptive models—after all, what good is a learner that isn't learning anything in particular—they are used quite regularly for data mining.

For example, the descriptive modeling task called pattern discovery is used to identify frequent associations within data. Pattern discovery is often used for market basket analysis on transactional purchase data. Here, the goal is to identify items that are frequently purchased together, such that the learned information can be used to refine the marketing tactics. For instance, if a retailer learns that swimming trunks are commonly purchased at the same time as sunscreen, the retailer might reposition the items more closely in the store, or run a promotion to "up-sell" customers on associated items.

Tip

Originally used only in retail contexts, pattern discovery is now starting to be used in quite innovative ways. For instance, it can be used to detect patterns of fraudulent behavior, screen for genetic defects, or prevent criminal activity.

The descriptive modeling task of dividing a dataset into homogeneous groups is called clustering. This is sometimes used for segmentation analysis that identifies groups of individuals with similar purchasing, donating, or demographic information so that advertising campaigns can be tailored to particular audiences. Although the machine is capable of identifying the groups, human intervention is required to interpret them. For example, given five different clusters of shoppers at a grocery store, the marketing team will need to understand the differences among the groups in order to create a promotion that best suits each group. However, this is almost certainly easier than trying to create a unique appeal for each customer.

Matching your data to an appropriate algorithm

The following table lists the general types of machine learning algorithms covered in this book, each of which may be implemented in several ways. Although this covers only some of the entire set of all machine learning algorithms, learning these methods will provide a sufficient foundation for making sense of other methods as you encounter them.

Model

Task

Chapter

Supervised Learning Algorithms

Nearest Neighbor

Classification

Chapter 3

naive Bayes

Classification

Chapter 4

Decision Trees

Classification

Chapter 5

Classification Rule Learners

Classification

Chapter 5

Linear Regression

Numeric prediction

Chapter 6

Regression Trees

Numeric prediction

Chapter 6

Model Trees

Numeric prediction

Chapter 6

Neural Networks

Dual use

Chapter 7

Support Vector Machines

Dual use

Chapter 7

Unsupervised Learning Algorithms

Association Rules

Pattern detection

Chapter 8

k-means Clustering

Clustering

Chapter 9

To match a learning task to a machine learning approach, you will need to begin with one of the four types of tasks: classification, numeric prediction, pattern detection, or clustering. Certain tasks make the choice of algorithm simpler. For instance, if you are undertaking pattern detection, you will likely employ association rules. Similarly, a clustering problem will likely utilize the k-means algorithm while numeric prediction will utilize regression analysis or regression trees.

For classification, more thought is needed to match a learning problem to an appropriate classifier. In these cases, it is helpful to consider the various distinctions among the algorithms. For instance, within classification problems, decision trees result in models that are readily understood, while the models of neural networks are notoriously difficult to interpret. If you were designing a credit-scoring model, this could be an important distinction because law often requires that the applicant must be notified about the reasons he or she was rejected for the loan. Even if the neural network was better at predicting loan defaults if the predictions cannot be explained, then it is useless.