Change for the Better: Improving Predictions by Automating Drift Detection

Drifting data poses three problems: detecting and assessing drift-related model performance degradation; generating a more accurate model from the new data; and deploying a new model into an existing machine-learning pipeline. Using a real-world predictive maintenance problem as an example, we demonstrate a solution that addresses each of these challenges. We reduce the complexity and costs of operating the system - as well as increase its reliability - by automating both drift detection and data labelling. After watching this video, you will understand how to develop streaming analytics on a desktop, deploy those solutions to the cloud, and apply AutoML strategies to keep your models up-to-date and their predictions as accurate as possible.

Recorded during Big Things Conference 2021