Machine Learning
Welcome! You can use this site to tell the Azure Machine Learning team what features you would like to see.
Azure Machine Learning enables developers and data scientists to build advanced analytic web services in minutes and hours, eliminating the heavy lifting associated with deploying machine learning in modern data driven applications.
-
Package management
The current way of adding additional R packages to Azure ML is insanely time consuming. There should be a way to install packages from CRAN along with the required dependencies.
I have tried to install the tidytext package using the method described here: https://blogs.msdn.microsoft.com/benjguin/2014/09/24/how-to-upload-an-r-package-to-azure-machine-learning/
When installing tidytext into a clean package library directory there are a total of 23 dependent packages that get installed. I then need to reference which ones are already installed on Azure, check if the version is up to date, package, upload and then install individually (in the correct order).
Azure ML Studio should be able…
55 votes -
More flexibility for Support Vector Machines
In other machine learning tools, SVM supports different kernel functions, and each function may have additional parameters. I assume "Project to the unit-sphere" is a kernel function, but could not find a definition for that function.
A kernel I found useful in many cases was the Radial Basis Function (RBF, parameters C or lambda, gamma), but there are others. It would also be useful to allow defining a custom kernel function, for example with an R script (or whatever language is most appropriate for the SVM implementation used)
55 votes -
More Help on Error 0000: Internal Error's
I get lots of these. A little bit more help finding out how to rectify them would be useful. Mainly on the training steps. I suspect its too many columns, but can't tell. Can more information be surfaced in the error logs?
50 votesstarted ·Adminmluservoice (Program Manager, Microsoft Azure Machine Learning, Microsoft Azure) responded
We continue to improve error. We have improved many errors and will continue to do more. This is ongoing activity for us
-
It would be great to add an association rules module to AML
Association rules models can be used in many situations:
- basket analysis (retail)
- finding a sequence of events that generate a failure (predictive maintenance)
- finding paths that conclude to a purchase (web mining analysis)
...AML should have this kind of algorithm (Apriori, CARMA...) that answer to many business situations.
Hope it will be coming soon!
Franck
40 votes -
Export Azure Machine Learning Studio's Workflow to Python/Anaconda codes.
I want to use the workflow at on premise Python/Anaconda environments for more speed, more customize, more cost-saving.
36 votes -
Bring "ML Studio" and "ML Services" together!
Give the ability to migrate existing "Azure ML Studio Webservices" to "Azure ML Model Management" OR provide the ability to deploy Models from Workbench to "Azure ML Studio Webservices"
36 votes -
Put a for loop container in Azure ML like in SSIS
In SSIS there is a loop container that allows to repeat some parts of the process as a for loop.
It would be nice to have the same functionality in Azure ML.36 votes -
Support CNTK on Azure ML
Using CNTK for RNN is great but making the models operational is a pain. It would be great to be able to use the CNTK libraries within the Azure ML service.
35 votes -
Custom Named entity recognition in Azure Machine Learning
Apart from identifying the name, location and organization, it would be useful for customers to define their custom named entities such as Email address, Contact number, date of birth or article date with year.
34 votes -
Within Azure ML Python, provide opencv library access.
With respect to Python in Azure ML, could opencv library access be provided? Currently, 'import cv2' fails. I am aware there are opencv modules and API in Azure, but I need to train over my own data which are not human faces.
33 votes -
Undo
Provide an undo feature to allow users to revert the previous change and series of changes (by repeatedly applying undo) within the interface. The existing Discard Changes isn't granular enough.
32 votes -
Allow to choose between SQLite and T-SQL in the Apply SQL Transformation module
It'll be a really good option write T-SQL scripts in the Apply SQL Transformation module since we are in a MS environment.
Thank you.
32 votes -
Multi Label Classification
Support for Multi Label Classification Algorithms using the Azure ML Tasks.
This is a very common request mostly in text analysis where one phrase can belong to one or more categories and we need to know which ones.
32 votes -
Automatic Integration of Application Insight Data
I would like to be able to connect the Application Insight Data from our WebApi directly into Machine Learning Experiment.
The goal for us to be able to predict things depending of the real time outcome30 votes -
Pre-install custom R packages
I want to expose an R package I wrote as a web service. Since it's a custom R package they docs say the install has to happen every time the user user my service.
There's going to be a lot of repetitive overhead going this route, Any chance I can pre-install my package so the user only has to consume using the web service?
30 votesplanned ·Adminmluservoice (Program Manager, Microsoft Azure Machine Learning, Microsoft Azure) responded
We are planning to release a fix pretty soon (in week or 2) to address this issue
-
Learning curve
Would be very nice to use a built-in Learning Curve plot which can be plotted after using the evaluate model module. This would require the sampledata to be handled by row and have the errors computed for the validation dataset and trainingdataset by row as well. The R2 value can be derived from the algorithm module that is selected and you could add another parameter input for theta. Would make it much easier to analyse if your algorithm is suffering from bias and variance.
30 votes -
dark theme
Can we have a dark theme for the workbench please?
30 votes -
Restore the last experiment deleted by mistake
It'd be nice if I could restore the last experiment I deleted by mistake.
Thank you
29 votes -
Add support for Asociation (Market Basket) Analysis
This is such a basic algorithm, I can't believe this is missing from ML
29 votes -
Upgrade Everything (R, Anaconda, IPython, etc.)
Just about every package is dated. IPython is using version 3.x.x but version 5 is available. Anaconda is using 2.x.x but version 4 is available. R is using 3.1 or Microsoft Open R and both of these are behind by several releases. Updating these would also update Pandas, Sklearn, etc.
26 votes
- Don't see your idea?