What is a baseline model and why we need it? Checkout here!

• The binary split (→ OneR classifier) shown in this fastai tutorial turns out to be one of the most effective classifiers in real-world datasets, compared to the algorithms in use in 1993. Since it's so simple and surprisingly effective, it makes for a great baseline -- that is, a starting point that you can use to compare your more sophisticated models to.
• Best start: Random Forest are good models to start with. Reason can be found here.
• Model mining für schnelle und erklärbare Modele.

<aside> 💡 A great baseline model is LightGBM

</aside>

The baseline model is the starting point. From there we start looking for the best hyperparameters:

Find best hyperparameter for your model

Applications

Creating a Baseline Regression Model

Creating a Baseline Classification Model

Scikit Learn Dummy Model ( →benchmark)

Baseline models with GXBoost:

XGBoost baseline model for regression

XGBoost baseline model for classification

Human Learn

A powerful tool from calmcode to use (or add) rules instead of complex ML models. This package contains scikit-learn compatible tools that should make it easier to construct and benchmark rule based systems that are designed by humans.

• documentation