Microsoft Azure Essentials Azure Machine Learning

by Jeff Barnes

Microsoft Azure Essentials from Microsoft Press is a series of free ebooks designed to help you advance your technical skills with Microsoft Azure.
 
This third ebook in the series introduces Microsoft Azure Machine Learning, a service that a developer can use to build predictive analytics models (using training datasets from a variety of data sources) and then easily deploy those models for consumption as cloud web services. The ebook presents an overview of modern data science theory and principles, the associated workflow, and then covers some of the more common machine learning algorithms in use today. It builds a variety of predictive analytics models using real world data, evaluates several different machine learning algorithms and modeling strategies, and then deploys the finished models as machine learning web services on Azure within a matter of minutes. The ebook also expands on a working Azure Machine Learning predictive model example to explore the types of client and server applications you can create to consume Azure Machine Learning web services.
 
Watch Microsoft Press’s blog and Twitter (@MicrosoftPress) to learn about other free ebooks in the Microsoft Azure Essentials series.

Genres: Programming, Computers

236 pages, Kindle Edition

First published April 25, 2015

About the author

A freelance writer and fifth-generation Nebraskan, Jeff Barnes is a former newspaper reporter and editor, past chairman of the Nebraska Hall of Fame Commission, and former marketing director for the Durham Western Heritage Museum. He traveled more than 13,000 miles in researching and photographing Forts of the Northern Plains, his first book.

Reviews

[George · Rating 3 out of 5]

This is a tutorial to Microsoft's Azure.

It goes through 3 experiments of common Machine Learning problems in order to introduce the user to the platform. Given it's a technical book about a product (a very interesting one) that should normally require a solid background on Machine Learning, I am just going to write down a set of remarks as I don't expect anyone to read the review in order to decide whether to get the book or not anyway.

Remark 1:
I believe that you can't address people as Data Scientists while at the same time you say things that actually imply that they don't even know what's going on and will likely use the Machine Learning algorithms through Azure as black boxes.

You are not a Data Scientist because you can use Azure, much in the same way that you are not a DJ because you can use Spotify.

Remark 2:
Why the hurry to publish such a book when the product is not yet in general availability status? This means that it will probably still have to undergo changes -possibly major ones- that will render the whole guide out of date.

Remark 3:
Seems that the book (and code) was written in a hurry and was not proof-read sufficiently.

Remark 4:
I would prefer to see classification, regression and clustering being mentioned and defined in high-level after, and not before defining supervised and unsupervised learning.

Remark 5:
Overall I believe that the importance of data preprocessing was not stressed enough. It also feels like Azure is not yet powerful enough with regards to preprocessing. For example, datasets aimed to be used in NLP experiments will need sophisticated preprocessing which I am not sure Azure can offer.

Things I liked (about Azure):
* I like the Visualise option a lot. It provides useful statistical information for the data effortlessly.
* The Designer and the way that you can see your whole experiment in a graphical representation is really great
* Publishing an ML experiment as a web service is extremely easy.
* The automatic workbook generation that is able to trigger api calls to the web service on its own is really impressive.
* The Cleaning feature with various modes for filling missing data is nice.

[Shai Sachs · Rating 3 out of 5]

This book is a pretty decent introduction to the basics of machine learning, although it doesn't go into much more depth than the three basic kinds of ML algorithms. That is more or less enough for this book, though; the real point is to illustrate how Azure ML works, and it does a reasonable job of walking you through the ML studio, and explaining how to use the APIs it produces. It does get a bit repetitive in the mechanics, which is annoying more than anything else. However, the book also doesn't do such a good job of exploring the more advanced uses of ML studio (like using custom R code in a module). That said, I did appreciate the dive into the recommendation engine, which seems like a great tool with a lot of interesting applications. All told - if you're trying to figure out how to get your feet wet with machine learning, this is definitely a good place to start.

[Johnny · Rating 3 out of 5]

Well written introduction into the capabilities of Azure for machine learning. The examples of machine learning are good to understand the problem you can solve with it, but far too short to really understand this topic as well. Therefore, don’t expect to learn machine learning as a by-product of understanding what Azure can offer you.

With all the changes in the Azure portal you shouldn’t spend too much time with learning the dialogues and screens. Chances are good that all that has changed even before you start reading this book. Nevertheless, this is a helpful book and can give you guidance in how use Azure for this kind of tasks.

[Yoly · Rating 5 out of 5]

Great introduction to the service.