Skip to main content
Documentation
< All Topics
Print

Data Drift Report 

The data drift detection is a way to evaluate whether the new incoming data is drifted from the data that was used in the model training. MLOps monitors for drift of every feature – both numerical and categorical features – and report them as graphs.  

Figure 4.8a shows a data drift detection report page. Clicking on each column will show detailed drift (as trend graph, Figure 4.8b). 

Figure 4.8a: Data Drift Detection report page. 

Figure 4.8b: Field specific drift trend. The points within the green band are compliant data while the data outsides the green area are drifted data. 

Table of Contents