Reader small image

You're reading from  Machine Learning Engineering with MLflow

Product typeBook
Published inAug 2021
PublisherPackt
ISBN-139781800560796
Edition1st Edition
Tools
Right arrow
Author (1)
Natu Lauchande
Natu Lauchande
author image
Natu Lauchande

Natu Lauchande is a principal data engineer in the fintech space currently tackling problems at the intersection of machine learning, data engineering, and distributed systems. He has worked in diverse industries, including biomedical/pharma research, cloud, fintech, and e-commerce/mobile. Along the way, he had the opportunity to be granted a patent (as co-inventor) in distributed systems, publish in a top academic journal, and contribute to open source software. He has also been very active as a speaker at machine learning/tech conferences and meetups.
Read more about Natu Lauchande

Right arrow

Monitoring data drift and model performance

In this section, we will run through an example that you can follow in the notebook available in the GitHub repository (at https://github.com/PacktPublishing/Machine-Learning-Engineering-with-MLflow/tree/master/Chapter11/model_performance_drifts) of the code of the package. We will run through the process of calculating different types of drift and exploring its integration with MLflow.

One emergent open source tool in the space of monitoring model performance is called Evidently (https://evidentlyai.com/). Evidently aids us in analyzing ML models during the production and validation phases. It generates handy reports integrated with pandas, JSON, and CSV. It allows us to monitor multiple drifts in ML models and their performance. The GitHub repository for Evidently is available at https://github.com/evidentlyai/evidently/.

In this section, we will explore the combination of Evidently with MLflow, in order to monitor data drift and...

lock icon
The rest of the page is locked
Previous PageNext Page
You have been reading a chapter from
Machine Learning Engineering with MLflow
Published in: Aug 2021Publisher: PacktISBN-13: 9781800560796

Author (1)

author image
Natu Lauchande

Natu Lauchande is a principal data engineer in the fintech space currently tackling problems at the intersection of machine learning, data engineering, and distributed systems. He has worked in diverse industries, including biomedical/pharma research, cloud, fintech, and e-commerce/mobile. Along the way, he had the opportunity to be granted a patent (as co-inventor) in distributed systems, publish in a top academic journal, and contribute to open source software. He has also been very active as a speaker at machine learning/tech conferences and meetups.
Read more about Natu Lauchande