Cloud Cover 151: Azure Machine Learning with Parmita Mehta

Sign in to queue

Description

In this episode Chris Risner and Haishi Bai are joined by Parmita Mehta, Program Manger on the Azure Machine Learning team.  Parmita gives an overview of the new Azure Cloud Machine Learning Service.  Using the Cloud ML service, Parmita demonstrates how you can create experiments to detect patterns in data.  In the example shown, Parmita uses Cloud ML to predict family income level based off of census data.  This is tested using modules pre-built by MS Research, Xbox, and Bing!  After building and testing out an experiment, Parmita shows how easy it is to push the experiment to production as a web service which can be used to run individual tests or batches.

Links from the show:

Like Cloud Cover on Facebook!



Follow @haishibai2010

Embed

Download

Download this episode

The Discussion

  • User profile image
    sokhaty

    Looks pretty cool. I'm a bit surprised that there is no code samples in F#. Is it possible to add 3rd party or custom-build components into the tool panel? For example, to add a database connector or custom missing value proxy?

  • User profile image
    Jan​Kratochvil

    I have to say, this is really impressive.

    I think a lot of companies who own data will be able to leverage this technology in very interesting ways.

    Great job!

  • User profile image
    Mid

    Very impressive work according to the video but not really practical in real live data analysis.

    that tool looks like an upgraded version of Excel.

    by the way her explanation sucks!

  • User profile image
    Chunde

    She used the Chrome browser!:-.

    Anyway, it's kind of interesting. But not very powerful

  • User profile image
    Barony

    @Chunde:Why do you think so?

    Found this rather confusing for a Machine Learning novice.
    Even Selected Columns for the different modules are not clearly defined Start with All columns or No Columns ? Clearing down Columns and Labels is not well explained
    The Classification modules do not tie up with
    I think I need to find a simpler Video to understand how to use this toolset.

Add Your 2 Cents