As a senior consult at Kohera, Belgium, I’m implementing the client Datawarehouse and Business Intelligence solutions. With 25 years of experience in various architectures and businesses, I’ve a broad view on the different data platforms used in the field. Without exceptions, all environments have their strength and can be combined, for this reason, I believe that hybrid-cloud architectures are the way to go. The last few years, my interest in advanced analytics has increased a lot. While this was evolving, the possibilities for operationalizing the data scientist work becomes available and make it more interesting than ever. Also experimenting with Azure Machine Learning and In-Database Machine learning technologies brings the search for the best data driven solutions on top of my activities. Sharing my experiences within this fields looks the next logic chapter in my career.
Supervised learning with Microsoft ML Services for in-database architecture
Data scientists conquering more and more territory in todays organizations. The insights they give have made a big contribution to the company’s success. The experimental and researching face is now shifting towards automated processes where business processes can benefit from this new knowledge. With Microsoft ML services we can build and integrate the data scientists work into our existing architecture directly on top of your database using the in-database integration facility.
In this session, I explain one possible solution for creating a supervised learning model on an existing SQL server 2017 installation, the automation for the model regeneration, retraining, rescoring, versioning of the model and finally the usage of it in your SQL code. Some architectural questions regarding the choice between Azure ML or In-Database implementation will be explained.