Glossary term
Glossary term
Evaluation and Benchmarks
A model used as a reference point for comparing how well another model (typically, a more complex one) is performing. For example, a logistic regression model might serve as a good baseline for a deep model.
For a particular problem, the baseline helps model developers quantify the minimal expected performance that a new model must achieve for the new model to be useful.
Created for this library
A pricing team always reports a heuristic baseline alongside the ML model so executives see the lift the model actually provides.
A search relevance team treats the previous quarter's ranker as the baseline and requires any candidate to beat it on the same evaluation set.
A fraud team uses a simple rules-based baseline as the comparison point for every new model so business sponsors can interpret model gains.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License