(deprecated) Web services examples using R code on Azure Machine Learning and published to Microsoft Azure Marketplace
The Microsoft DataMarket is being retired and these APIs have been deprecated.
You can find many useful example experiments and APIs in the Cortana Intelligence Gallery. For more information about the Gallery, see Share and discover resources in the Cortana Intelligence Gallery.
In this article are example web services were created using Azure Machine Learning and published to the Azure Marketplace. Each web service example has a comprehensive document attached, embedding sample data sets for testing the services and explaining how the user can create a similar service themselves.
In Azure Machine Learning Studio, users can write R code and with a few clicks, publish it as a web service for applications and devices to consume around the world.
Try Azure Machine Learning for free
No credit card or Azure subscription needed. Get started now >
From producing simple calculators that provide statistical functionality to creating a custom text-mining sentiment analysis predictor, both new and experienced R users can benefit from the ease with which users of Azure Machine Learning can operationalize R code. While these web services could be consumed by users, potentially through a mobile app or a website, the purpose of these web services examples is to show how Machine Learning can operationalize R scripts for analytical purposes and be used to create web services on top of R code.
Each example includes a C# example for web service consumption.
Consider the following scenarios.
Scenario 1: Generic model
A user works with a generic model that can be applied to a new user’s data, such as a basic forecasting of time series data or a custom-built R method with advanced analytics. This user publishes the model as a web service for others to consume with their data.
- Binary Classifier
- Cluster Model
- Multivariate Linear Regression
- Forecasting - Exponential Smoothing
- ETS + STL Forecasting
- Forecasting - Autoregressive Integrated Moving Average (ARIMA)
- Survival Analysis
Scenario 2: Trained model – specific data
A user has data that provides useful predictions through R code, such as a large sample of personality questionnaires clustered through a k-means algorithm to predict the user’s personality type, or health survey data that can be used to predict an individual’s risk for lung cancer with a survival analysis R package. The user publishes the data through a web service that predicts a new user’s outcome.
Scenario 3: Trained model – generic data
A user has generic data (such as a text corpus) that allows a web service to be built and applied generically across different types of use cases and scenarios.
Scenario 4: Advanced calculator
A user provides advanced calculations or simulations that don’t require any trained model or fitting of a model to the user’s data.
For frequently asked questions on consumption of the web service or publishing to the Marketplace, see here.