Introduction to Microsoft Machine Learning Studio

I wrote an article titled “Introduction to Machine Learning Studio” in the September 2014 issue of MSDN Magazine. See Actually the title is a bit inaccurate (on purpose). Machine Learning Studio (ML Studio) is a Web application that allows you to create a machine learning system using drag-and-drop modules — no coding at all. ML Studio is essentially the UI for Microsoft Azure Machine Learning, which is a Web service that does all the work.

Figure1-A Complete ML Studio Experiment

In the article, I explain, step-by-step, how to create a system that predicts whether a U.S. Congressman is a Republican or Democrat, based on their voting record. The example uses Logistic Regression, but ML Studio supports several different ML classification techniques, including neural networks.

Overall, I am quite a fan of ML Studio. Using the tool saves a ton of time compared to my normal code-from-scratch approach. On the downside, ML Studio does not yet have an SDK so it’s not currently possible to create custom-coded modules using C# (although ML Studio can execute R language scripts, which is sort of a customization approach).

And, I’m really not a fan of Cloud-based applications. I know they’re the wave of the future and inevitable, but I prefer having my entire system local, sitting in a PC in my office. I get annoyed by Cloud-based systems that are constantly changing — I like a stable platform to work on even if it means I need to install upgrades every now and then.

This entry was posted in Machine Learning. Bookmark the permalink.