Most model assessment metrics, such as Lift, AUC, KS, ASE, require the presence of the target/label to be in the data. This is always the case at the time of model training. But how can I ensure that the developed model can be applied to new data for prediction?
Tag: solving the last mile
Introducing the feature contribution index for model assessment
Creating a model validation pipeline
This post describes a fully automated validation pipeline for analytical models as part of an analytical platform, which has been set up recently as part of a customer project.
Model validation testing in the age of DevOps
Let us now take a look at a well-known metaphor for test case development in the software industry. We are referring to the idea of the “test pyramid."
A structured approach to model validation
In total, there are four posts in this blog series, this is the first post describing some basic principles of the DevOps (or ModelOps) approach.