Data Drift Explainability: Interpretable Shift Detection with NannyML | by Marco Cerliani | Jun, 2022
Alerting Meaningful Multivariate Drift and ensuring Data QualityPhoto by FLY:D on UnsplashModel monitoring is becoming a hot trend in machine learning. With the crescent hype in the activities concerning the MLOps, we register the rise of tools and research about the topic.One of the most interesting is for sure the Confidence-based Performance Estimation (CBPE) algorithm developed by NannyML. They implemented a novel procedure to estimate future models' performance degradation in absence of ground truth. It may yield…