From the course: Machine Learning with SageMaker by Pearson

Unlock this course with a free trial

Join today to access over 25,200 courses taught by industry experts.

Using SageMaker Model Monitor for data drift and quality

Using SageMaker Model Monitor for data drift and quality - Amazon SageMaker Tutorial

From the course: Machine Learning with SageMaker by Pearson

Start my 1-month free trial Buy for my team

Using SageMaker Model Monitor for data drift and quality

“

Once we have our models deployed in production, we need a way to monitor them over time in order to detect drift, as well as a change in accuracy or quality of our model while it's running in production. SageMaker Model Monitor is a tool to do that. It gives us continuous monitoring of the data as well as predictions for our model. And it enables automated alerting for deviations from those baselines that are determined post-training of our model. So why do we care? Well, we need to ensure that our input data remains consistent with the training data. So we're trained on this large data set, and as we're feeding information in in order to receive predictions, we need to ensure that the quality of those predictions aligns with the quality that was measured after we created our model. This can identify a shift in distribution, that being data drift, that can impact our performance, can detect missing or corrupted data in real time, maintains compliance and reliability of our models…

Contents