This is a reference implementation of Microsoft's Matchbox Recommender provided with ready-to-use Azure Machine Learning Studio experiment. This Recommender is then deployed as Azure Machine Learning Studio web service and requested via Azure Api Management. Simple web application written in React, which requests mentioned recommender, is also included.
- Open this experiment in Azure Machine Learning Studio
- Run & Publish as web service
- I strongly advise to train the model on MovieLens 20M dataset as I did in the demo above
- Run & Publish as web service
- Create Azure API Management service
- Create endpoint calling your Azure Machine Learning Studio published experiment
- Clone & install frontend application from this repo
- Fill in correct API & subscription keys inside
./frontend/.env(follow./frontend/.env_example)
- Fill in correct API & subscription keys inside
- Run frontend application
- Train the model yourself Azure Machine Learning Studio experiment
- Use MovieLens 20M dataset for greater accuracy