Actuarial Outpost
 
Go Back   Actuarial Outpost > Actuarial Discussion Forum > Software & Technology
FlashChat Actuarial Discussion Preliminary Exams CAS/SOA Exams Cyberchat Around the World Suggestions


Upload your resume securely at https://www.dwsimpson.com
to be contacted when our jobs meet your skills and objectives.


Reply
 
Thread Tools Search this Thread Display Modes
  #1  
Old 06-25-2018, 05:31 PM
paulbernal paulbernal is offline
Member
 
Join Date: Oct 2009
Posts: 240
Default Automating Azure ML Experiments Execution

Dear friends,

I created a forecasting model in R using Microsoft Azure Machine Learning Studio. However, I would like to automate the execution of the experiments so that I dont have to manually sign in Azure ML Studio and hit the run button.

Anyone here has any idea of how to automate azure ml experiment execution?

Any help and/or guidance will be greatly appreciated,

Best regards,

Paul
Reply With Quote
  #2  
Old 06-26-2018, 12:37 PM
whoanonstop's Avatar
whoanonstop whoanonstop is offline
Member
Non-Actuary
 
Join Date: Aug 2013
Location: Los Angeles, CA
Studying for Spark / Scala
College: College of William and Mary
Favorite beer: Orange Juice
Posts: 5,899
Blog Entries: 1
Default

Quote:
Originally Posted by paulbernal View Post
Dear friends,

I created a forecasting model in R using Microsoft Azure Machine Learning Studio. However, I would like to automate the execution of the experiments so that I dont have to manually sign in Azure ML Studio and hit the run button.

Anyone here has any idea of how to automate azure ml experiment execution?

Any help and/or guidance will be greatly appreciated,

Best regards,

Paul
So I don't use Microsoft Azure ML (although I've seen the GUI with the drag and drop models and it makes me cringe), but I might be able to help.

I guess my first question is if you're writing it in R, why does it have to be in Azure? You should be able to connect to the databases, fit your model, and predict all from R, which should eliminate your need to connect to Azure.

If it must remain in Azure, surely they have some scheduling service for you to update/run at any given time or in specific increments. I did a quick search and found this: https://azure.microsoft.com/en-us/services/scheduler/

Again, not sure how the product structure works for Azure, but if you have access to this then I'm sure the answer lies there and you should investigate more. Otherwise, you could write something that signs into Azure for you and clicks the go button, but that feels inefficient if the scheduler above exists for you.

-Riley
__________________
Reply With Quote
  #3  
Old 06-26-2018, 04:27 PM
glassy glassy is offline
Member
Non-Actuary
 
Join Date: May 2015
Posts: 3,889
Default

How long does the process of logging and and clicking run actually take you?
Reply With Quote
  #4  
Old 06-26-2018, 05:55 PM
paulbernal paulbernal is offline
Member
 
Join Date: Oct 2009
Posts: 240
Default

15 mins approx. The thing is that I want to operationalize the whole experiment. I need to create a job that fires execution of the experiment with a certain frequency (once every 24 hours, once a month, once every 3 months, etc.).

My goal is to automate execution of the Azure ML model/experiment.
Reply With Quote
  #5  
Old 06-26-2018, 05:56 PM
paulbernal paulbernal is offline
Member
 
Join Date: Oct 2009
Posts: 240
Default

Quote:
Originally Posted by glassy View Post
How long does the process of logging and and clicking run actually take you?
15 mins approx. The thing is that I want to operationalize the whole experiment. I need to create a job that fires execution of the experiment with a certain frequency (once every 24 hours, once a month, once every 3 months, etc.).

My goal is to automate execution of the Azure ML model/experiment.
Reply With Quote
  #6  
Old 06-27-2018, 07:14 AM
kevinykuo kevinykuo is offline
CAS
 
Join Date: Nov 2017
Posts: 24
Default

You could have an Azure CLI script that runs the experiment and try to automate that, e.g. https://stackoverflow.com/questions/...ure-cli-script

Why do you need ML studio for what you're doing? In other words, does it accomplish something that running R on Azure doesn't?
Reply With Quote
  #7  
Old 06-27-2018, 09:54 AM
paulbernal paulbernal is offline
Member
 
Join Date: Oct 2009
Posts: 240
Default

Quote:
Originally Posted by kevinykuo View Post
You could have an Azure CLI script that runs the experiment and try to automate that, e.g. https://stackoverflow.com/questions/...ure-cli-script

Why do you need ML studio for what you're doing? In other words, does it accomplish something that running R on Azure doesn't?
Dear friend Kevinykuo,

I could do the exact same thing in R, or R Studio or directly in Azure ML Studio. The company, however, has lots of Microsoft Applications (Azure SQL Server, Azure ML Studio, Power BI, etc.).

I can go ahead and run what I have directly in Azure ML Studio, however, what I want to do is to create a job that executes my experiments on a certain schedule, without me having to sign in Azure ML and manually hitting the Run button.

I want to make experiment execution automatic, I want to create some sort of activity that triggers the automatic execution of existing experiments, why? Because we are working on some other projects that require this to operate that way.
Reply With Quote
Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off


All times are GMT -4. The time now is 08:35 PM.


Powered by vBulletin®
Copyright ©2000 - 2018, Jelsoft Enterprises Ltd.
*PLEASE NOTE: Posts are not checked for accuracy, and do not
represent the views of the Actuarial Outpost or its sponsors.
Page generated in 0.21359 seconds with 9 queries