A hands-on discussion of machine learning with Mahout.
Mahout in Action is a hands-on introduction to machine learning with Apache Mahout. Following real-world examples, the book presents practical use cases and then illustrates how Mahout can be applied to solve them. Includes a free audio- and video-enhanced ebook.
preface
acknowledgments
about this book
about multimedia extras
about the cover illustration
Part 1 Recommendations
1. Chapter 2 Introducing recommenders
1.1. Defining recommendation
1.2. Running a first recommender engine
1.3. Evaluating a recommender
1.4. Evaluating precision and recall
1.5. Evaluating the GroupLens data set
1.6. Summary
2. Chapter 3 Representing recommender data
2.1. Representing preference data
2.2. In-memory DataModels
2.3. Coping without preference values
2.4. Summary
3. Chapter 4 Making recommendations
3.1. Understanding user-based recommendation
3.2. Exploring the user-based recommender
3.3. Exploring similarity metrics
3.4. Item-based recommendation
3.5. Slope-one recommender
3.6. New and experimental recommenders
3.7. Comparison to other recommenders
3.8. Summary
4. Chapter 5 Taking recommenders to production
4.1. Analyzing example data from a dating site
4.2. Finding an effective recommender
4.3. Injecting domain-specific information
4.4. Recommending to anonymous users
4.5. Creating a web-enabled recommender
4.6. Updating and monitoring the recommender
4.7. Summary
5. Chapter 6 Distributing recommendation computations
5.1. Analyzing the Wikipedia data set
5.2. Designing a distributed item-based algorithm
5.3. Implementing a distributed algorithm with MapReduce
5.4. Running MapReduces with Hadoop
5.5. Pseudo-distributing a recommender
5.6. Looking beyond first steps with recommendations
5.7. Summary
Part 2 Clustering
6. Chapter 7 Introduction to clustering
6.1. Clustering basics
6.2. Measuring the similarity of items
6.3. Hello World: running a simple clustering example
6.4. Exploring distance measures
6.5. Hello World again! Trying out various distance measures
6.6. Summary
7. Chapter 8 Representing data
7.1. Visualizing vectors
7.2. Representing text documents as vectors
7.3. Generating vectors from documents
7.4. Improving quality of vectors using normalization
7.5. Summary
8. Chapter 9 Clustering algorithms in Mahout
8.1. K-means clustering
8.2. Beyond k-means: an overview of clustering techniques
8.3. Fuzzy k-means clustering
8.4. Model-based clustering
8.5. Topic modeling using latent Dirichlet allocation (LDA)
8.6. Summary
9. Chapter 10 Evaluating and improving clustering quality
9.1. Inspecting clustering output
9.2. Analyzing clustering output
9.3. Improving clustering quality
9.4. Summary
10. Chapter 11 Taking clustering to production
10.1. Quick-start tutorial for running clustering on Hadoop
10.2. Tuning clustering performance
10.3. Batch and online clustering
10.4. Summary
11. Chapter 12 Real-world applications of clustering
11.1. Finding similar users on Twitter
11.2. Suggesting tags for artists on Last.fm
11.3. Analyzing the Stack Overflow data set
11.4. Summary
Part 3 Classification
12. Chapter 13 Introduction to classification
12.1. Why use Mahout for classification?
12.2. The fundamentals of classification systems
12.3. How classification works
12.4. Work flow in a typical classification project
12.5. Step-by-step simple classification example
12.6. Summary
13. Chapter 14 Training a classifier
13.1. Extracting features to build a Mahout classifier
13.2. Preprocessing raw data into classifiable data
13.3. Converting classifiable data into vectors
13.4. Classifying the 20 newsgroups data set with SGD
13.5. Choosing an algorithm to train the classifier
13.6. Classifying the 20 newsgroups data with naive Bayes
13.7. Summary
14. Chapter 15 Evaluating and tuning a classifier
14.1. Classifier evaluation in Mahout
14.2. The classifier evaluation API
14.3. When classifiers go bad
14.4. Tuning for better performance
14.5. Summary
15. Chapter 16 Deploying a classifier
15.1. Process for deployment in huge systems
15.2. Determining scale and speed requirements
15.3. Building a training pipeline for large systems
15.4. Integrating a Mahout classifier
15.5. Example: a Thrift-based classification server
15.6. Summary
16. Chapter 17 Case study: Shop It To Me
16.1. Why Shop It To Me chose Mahout
16.2. General structure of the email marketing system
16.3. Training the model
16.4. Speeding up classification
16.5. Summary
Appendix A: JVM tuning
Appendix B: Mahout math
Appendix C: Resources
index
© 2014 Manning Publications Co.
About the Technology
A computer system that learns and adapts as it collects data can be really powerful. Mahout, Apache's open source machine learning project, captures the core algorithms of recommendation systems, classification, and clustering in ready-to-use, scalable libraries. With Mahout, you can immediately apply to your own projects the machine learning techniques that drive Amazon, Netflix, and others.
About the book
This book covers machine learning using Apache Mahout. Based on experience with real-world applications, it introduces practical use cases and illustrates how Mahout can be applied to solve them. It places particular focus on issues of scalability and how to apply these techniques against large data sets using the Apache Hadoop framework.
This book is written for developers familiar with Java. No prior experience with Mahout is assumed.
What's inside
- Use group data to make individual recommendations
- Find logical clusters within your data
- Filter and refine with on-the-fly classification
- Free audio and video extras
About the reader
This book is written for developers familiar with Java. No prior experience with Mahout is assumed.
About the authors
Sean Owen helped build Google's Mobile Web search and launched the Taste framework, now part of Mahout. Robin Anil contributed the Bayes classifier and frequent pattern mining implementations to Mahout. Ted Dunning contributed to the Mahout clustering, classification, and matrix decomposition algorithms. Ellen Friedman is an experienced writer with a doctorate in biochemistry.
- combo $44.99 pBook + eBook
- eBook $35.99 pdf + ePub + kindle
FREE domestic shipping on three or more pBooks
The writing makes a complex topic easy to understand.
Essential Mahout, authored by the core developer team.
Dramatically reduces the learning curve.
Recommendations, clustering, and classification all lucidly explained.