<aside> 💡
Behind every decision powered by an ML model is a real human that is affected by it.
You’ll make a lot more headway by getting out of your bubble and spending time with those people than you will by running yet another hyperparameter sweep.
</aside>
Schnell gute und sogar erklärbare Models bauen. Das geht mit einem Trick wie hier beschrieben. The tool is included in the human-learn package from calmcode (see tutorial). More details in the documentation.
Overview about most popular ML-models for interpretation → here