Sunday, February 6, 2022

Amazon SageMaker Canvas

The two biggest trends in no-code right now are office automation and machine learning applications. The ability to predict business outcomes is extremely important to any organization and the ability to build the models to generate those predictions is driving the surge in no-code machine learning application development. 

Recently, Amazon released SageMaker Canvas, its new platform that's designed to allow every decision-maker in an organization to create predictive models based on the data they work with every day, without the need to be a data scientist. SageMaker Canvas lets the user gather data from different sources on-premises or online, choose which datasets to import into Canvas, pick the columns to use as input and correct any missing or inaccurate records. Once that's done Canvas uses the imported data to generate and test multiple predictive models and recommends the best performing model to the user. Users can select to make changes as far as which dataset columns to use for training and Canvas will continue to run tests until a final model is chosen.

As of last November, SageMaker Canvas was only available in certain areas in the U.S. and Europe but that should change in the near future. You can see an actual  example of how to use Canvas in making shipping predictions at: 

No comments:

Post a Comment