5/26/2015

Intro to Azure Machine Learning and R Language

I'm investigating the statistical language "R".  And why is that?  Well, being in the Microsoft Business Intelligence stack, you may have heard that they've acquired Revolution Analytics.

R is one of the preferred languages for statisticians, data miners and analytic professionals.  Now it will be baked into the Microsoft Business Intelligence stack.

I found a great post by leading Microsoft BI professional Jen Stirrup explaining the acquisition of Revolution Analytics.

I found a course on Pluralsight for "R Programming Fundamentals" and listened to the course.

I dabbled with Matlab and Octave and posted a blog about it last year.

I don't claim to be a Statistician or Data Scientist, but you have to start somewhere.

This morning, I saw a good tweet: "Why I Left My Data Science Master's Program" by @CPensive on @LinkedIn https://www.linkedin.com/pulse/why-i-left-my-masters-program-charles-pensig-1


The author provides some real world expectations on what to learn and how to spend your time wisely.

Here's an older article on running R in Microsoft Azure.

I did a sample Azure Machine Learning blog post in August 2014, posted here.

Recently, I got access to the Microsoft Machine Learning portal:


So, now to investigate.  When logging on, you get:

There's a bunch of samples provided in the Gallery:


Going to the Gallery, search for R by "Tags" we get,


Selecting a random Experiment, we get a good summary,


Open in Studio, it copies the project to the work space,


And we see R options,


The canvas shows the flow of steps, as each step is connected to other steps, performing various functions along the way,


It added some entries to the My Datasets,


You can see the same files were added to the Dataset screen,


Back on the Experiment canvas, we selecting one of the tasks, we see the code in the editor to the right,


Scanning the code, there's some logic going on within the script,


Let's go ahead and run the Experiment,


Queued and running,


You can see each task running, when complete, is shows a green check,



Finished,


Viewing the Run History,



The documentation is provided for this experiment here.

There's an MSDN Forum to discuss Azure Machine Learning: https://social.msdn.microsoft.com/forums/azure/en-US/home?forum=MachineLearning

Here's a good blog source: http://www.r-bloggers.com/tag/rblogs/

As you can see, this portal is separate from the Azure website, which has Machine Learning option,


This post is by no means a deep dive into Azure Machine Learning or R language, but it does show you how easy the platform is to get started, load a sample project, run and see how an experiment is structured with actual code.

Revolution R Open – Latest Update Now Available for Free Download: http://blogs.technet.com/b/machinelearning/archive/2015/05/22/revolution-r-open-latest-update-now-available-for-free-download.aspx

Thanks for reading~!

No comments:

Post a Comment

We Interrupt this Broadcast