Microsoft has announced the General Availability of Azure Machine Learning service, a cloud-based predictive analytics service.

Microsoft hopes the service, which we first reported in June 2014 when it was launched as a preview for Microsoft partners, will simplify the process of using machine learning techniques to analyze large sets of data in the cloud, without needing as much input from data scientists. 

On his blog, Scott Guthrie says the browser-based machine learning IDE will let data scientists and develops create and automate machine learning workflows, adding:

“you can literally drag/drop hundreds of existing ML libraries to jump-start your predictive analytics solutions, and then optionally add your own custom R and Python scripts to extend them. Our Machine Learning IDE works in any browser and enables you to rapidly develop and iterate on solutions.”

The addition of Python makes the service more interesting, given the widespread use of Python for data analysis.

The new release lets you discover and create web services, train/retrain your models through APIs, manage endpoints and scale web services on a per customer basis. The release adds some new features, including the ability to create a configurable custom R module, incorporate your own train/predict R-scripts, and add python scripts using a large ecosystem of libraries such as numpy, scipy, pandas, and scikit-learn. You can now train on terabytes of data using “Learning with Counts”, use PCA or one-class SVM for anomaly detection, and modify, filter, and clean data using SQLite.

Another improvement is the Azure ML Community Gallery, which lets you:

“discover & learn experiments, and share through Twitter and LinkedIn.”

The term experiment is used to refer to an individual analysis. For developers, the fact marketplace apps are available to purchase through an Azure subscription may point to opportunities to create apps for sale. There are currently around twenty starter apps available.

There’s a new step-by-step guide (https://studio.azureml.net/) showing how to go from raw data to a consumable web service, and Microsoft has added the ability to use tools such as iPython Notebook and Python Tools for Visual Studio along with Azure ML.

If you’ve been trying the preview of Azure ML, the current preview pricing is in effect through March 31, 2015. Starting April 1, 2015, all new and existing Azure Machine Learning subscriptions will automatically be converted to the Standard version and billed accordingly; no action is needed by the user. If you don’t want to migrate you need to delete your workspaces before April 1, 2015. There is no change or charge for the Free version of Azure Machine Learning available from the Azure website.  

Azure Machine Learning service

Related Articles

Comments

Make a Comment or View Existing Comments Using Disqus

or email your comment to: comments@i-programmer.info