<aside> 💡 Experiments done with a bad evaluation strategy are a waste of time.

Conversely, a good evaluation setup is the most powerful tool for iterating to a good solution.

This is arguably the highest-leverage time investment in an ML project.

</aside>

<aside> 💡 First step when you construct your first model on a new problem

→ make sure your model can overfit!

</aside>

<aside> 💡 It is very important to Create a Baseline Model

</aside>

Important rules about evaluation of models 🤝

• Model evaluation is the best time saver → explanation here
• Rapid iteration → explanation here
• Pick 1 model and max it out → explanation here
• Understand you model weakness → explanation here

In real life macht es Sinn neben der Optimierung der Metric auch noch weitere Parameter zu beachten:

• Zeitaufwand
• Anzahl Features
• Special hardware requirements
• Latency of prediction process
• …

Videos zu Metrics