Written by Nikhil Joglekar, Miguel Fierro, and Le Zhang of AzureCAT. Edited by Nanette Ray. Published by Adam Boeglin from Microsoft patterns & practices.
This example scenario shows how a business can use machine learning to automate product recommendations for their customers. An Azure Data Science Virtual Machine (DSVM) is used to train a model on Azure that recommends movies to users based on ratings given to movies.
Recommendations are useful in various industries from retail to news to media. Potential applications include providing product recommendations in a virtual store, providing news or post recommendations, or providing music recommendations. Traditionally, businesses hired and trained assistants to make personalized recommendations to customers. Today, we can provide customized recommendations at scale by utilizing Azure to train models to understand customer preferences.
Components included in this example scenario:
- Data Science Virtual Machine (DSVM)
- Azure Blob storage
- Azure Machine Learning service
- Azure Cosmos DB
- Azure Container Instances
Topics covered include:
- Relevant use cases
- Architecture
- Components
- Alternatives
- Considerations
- Availability
- Scalability
- Security
- Deploy this scenario
- Related resources
Check out Build a real-time recommendation API on Azure for an in-depth guide on building and scaling a recommender service. Tutorials and examples of recommendation systems are available in the Microsoft Recommenders repository. There's also a growing list of example scenarios on the Azure Architecture Center.
AzureCAT Guidance
"Hands-on solutions, with our heads in the Cloud!"