Overfitting
What does overfitting mean?
Overfitting or overadaptation is a problem in machine learning where an AI model is too well adapted to specific training data.
The AI model therefore learns not only the essential patterns and correlations, but also specific characteristics of the training data (such as noise), which leads to poor performance with new, unknown data.
Figure legend:
Left: Underfitting
Right: Overfitting
