Implementing a monitoring workflow
Throughout the course, you've learned about the monitoring workflow. The first step is performance monitoring. If there are negative changes, the next steps involve multivariate drift detection to identify if drift caused the performance drop, followed by univariate drift detection to pinpoint the cause in individual features. Once the investigation results are in, you can take steps to resolve the issue.
To solidify this knowledge, in the exercise, you'll apply this process to the US Consensus dataset. The reference and analysis datasets are pre-loaded, and you have access to the CBPE estimator, uv_calc univariate calculator, and an alert_count_ranker for feature drift ranking.
This exercise is part of the course
Monitoring Machine Learning in Python
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Estimate the performance
estimator.____(____)
estimated_results = estimator.____(____)
estimated_results.____.____